ArticlePDF Available

Characterization of Nomex honeycomb core constituent material mechanical properties

Authors:
Manuscript preprint
Characterization of Nomex honeycomb core constituent material
mechanical properties
Rene Roya, Sung-Jun Parka, Jin-Hwe Kweona*, Jin-Ho Choib
a Department of Aerospace and System Engineering, Research Center for Aircraft Parts Technology,
Gyeongsang National University, 900 Gajwa-dong, Jinju, Gyeongnam, 660-701, Republic of Korea
b School of Mechanical Engineering, Research Center for Aircraft Parts Technology, Gyeongsang
National University, 900 Gajwa-dong, Jinju, Gyeongnam, 660-701, Republic of Korea
E-mails: rroy@gnu.ac.kr (R Roy), isungjun@nate.com (SJ Park), jhkweon@gnu.ac.kr* (JH Kweon),
choi@gnu.ac.kr (JH Choi). (* corresponding author)
Composite Structures
, Volume 117, November 2014, Pages 255-266
https://doi.org/10.1016/j.compstruct.2014.06.033
Abstract
Nomex honeycomb cores have been widely used in composite sandwich panels. To accomplish
meso-scale finite element modeling of these cores, cell wall mechanical properties are required, for
which limited data are available. In this work, tensile testing was performed on Nomex paper,
phenolic resin, and Nomex paper coated with phenolic resin. Flatwise tension and compression tests
were also performed on two types of Nomex honeycomb cores. Test results were calibrated in finite
element modeling to account for strain gage local stiffening effects and thickness normalization.
Identified cell wall material properties were implemented in a honeycomb core finite element model
and further calibrated by matching simulation results to manufacturer test data. The cells’ double-
wall thickness was also adjusted. These calibrations and adjustments led to an exact simulation
match with test data. Numerically matching cell wall material properties depends on modeling
sophistication and is subject to core test result variability and core construction differences among
manufacturers.
Keywords: Nomex honeycomb; Nomex paper; Phenolic resin; Mechanical properties; Finite element
analysis (FEA).
1. Introduction
Polymer composite sandwich construction is an attractive design option, as it typically offers an
excellent stiffness-to-weight ratio. Honeycomb type cores used in sandwich construction are among
the best performers in this regard and have been used extensively. Nomex (E.I. du Pont de Nemours
Corp., Wilmington, DE, USA) honeycomb cores can be a valuable choice given their flammability
properties, dielectric properties, environmental resistance and galvanic compatibility with face
materials. Nomex honeycomb cores are composed of adhered strips of Nomex paper dipped in
phenolic resin. Nomex paper is made from Nomex fibers, a meta-aramid chemical component.
Manuscript preprint
Nomex paper strips are typically adhered by bands of thermoset epoxy adhesive. This assembly,
once expanded to form honeycomb cells, is dipped into phenolic resin and cured to form the final
core. These cores are sold in different cell sizes (1.6 mm to 19 mm) and core densities (29 kg/m3 to
144 kg/m3) and can be ordered with a specific core height. Core manufacturers provide mechanical
properties for the core block, namely compression strength and modulus, and main direction shear
strength and modulus. Core tensile strength is often evaluated from a fabricated sandwich with a
flatwise tensile test [1]. This test can also serve to evaluate face-to-core bond quality [2]. A designer
using a Nomex honeycomb core can therefore consider these mechanical properties in his
calculations. Sandwich in-plane compression instabilities, namely panel buckling, intra-cell buckling
and skin wrinkling, must also be verified and analytical formulas are available to do so [3-4]. For
honeycomb core applications involving phenomena such as impact loading [5-7], concentrated
loading [8], inserts in the sandwich [9-11], or where the vibration properties of the sandwich are of
interest, experimental testing or finite element modeling (FEM) analysis may be desirable. FEM
models of honeycomb cores can rely on equivalent 3D solid elements, where a block of solid
elements is made to have the same properties as the core (compression, tension, and shear).
Honeycomb core failure may happen through cell material fracture or cell wall buckling, depending
on the loading regime and the core configuration [12-13]. Generally speaking, if the cell wall is thin
and long, it may be prone to buckling, whereas if it is thick and short, it may reach its critical failure
stress before buckling. In the case of cell wall buckling, this phenomenon cannot be represented
directly with an equivalent solid element FEM model. Recently, several researchers proceeded to
model the core with its detailed geometry using shell elements (meso-scale modeling) [6, 8, 13-17].
In this case, it is also desirable that the material properties used for shell elements are such that
the core’s global properties are reproduced (namely the core’s compression, tension and shear
moduli). A wide range of material property values have been used in the literature for meso -scale
FEM modeling of Nomex honeycomb cores [7-8, 13, 20]. In the references listed, the Young’s
modulus (E) used for honeycomb shell elements in the core thickness direction ranged from 1.9 to
3.5 GPa for different core size models (cell size, density). Apart from these material properties,
different modeling choices are possible, including basic cell geometry [4, 21], geometrical and
property variability within the cell (imperfections) [13, 15], or sandwich core/face boundary
conditions [14, 17]. It can be argued that modeling parameters such as accurate cell wall thickness,
cell geometry, and incorporation of imperfections have a non-negligible influence and will therefore
dictate the cell wall material property required to match a given core’s global properties. Material
test data of the cell wall material, which is essentially Nomex paper coated with phenolic resin, is
somewhat rare and is not provided by core manufacturers. Tsujii et al. [22] conducted tensile tests
on Nomex paper coated with phenolic resin, although the study is published in Japanese. Foo et al.
[23] tested the tensile properties of plain Nomex paper and measured tensile moduli of 3.40 GPa
(machine direction) and 2.46 GPa (cross direction). Staal [4] chose to perform miniature three point
flexural tests on rectangular cell wall sections cut out of a core. He obtained an average flexural
modulus of 3.81 GPa (in the core’s thickness direction) but found this approach sensitive to error in
the thickness measurement of the specimens. In this paper, experimental mechanical tests on Nomex
Manuscript preprint
paper, phenolic resin and phenolic resin-coated Nomex paper are reported. Characterizations of
Nomex honeycomb core density and geometry are also presented. Flatwise tension and
compression tests are also performed on Nomex honeycomb cores installed with strain gages.
Where applicable, the mechanical property results obtained are calibrated through FEM to account
for strain gage local stiffening effects and extra phenolic coating on the specimens tested compared
to a typical honeycomb material. These calibrated mechanical properties are then used in a meso-
scale shell element FEM model of a core, and the resulting simulated core properties are compared
to manufacturer test data and discussed.
2. Test procedure
2.1 Nomex paper tension test
Tensile tests were performed on Nomex type 410 paper with a nominal thickness of 0.05 mm [24].
A universal testing press with a 10kN load cell was used (model LR 10K from Lloyd Instruments Ltd),
as was the case for all load tests in this paper. A constant head displacement rate of 25 mm/min
was applied. The overall dimensions of the specimen were 50 mm (width) × 350 mm (length). Tabs
consisting of epoxy-adhered 25 mm × 50 mm #100 sandpaper sheets were used, giving an effective
gage length of 300 mm. Tests were performed in the paper roll direction (0°), transverse direction
(90°), and 45° for five specimens each. Strain data were derived by dividing the machine head
displacement by the initial gage length. This approach may be vulnerable to grip slippage and
machine compliance affecting the displacement data. However, the force level in the tests remained
relatively low (maximum force range 92187 N), so we assumed that machine compliance is small.
Additionally, no grip slippage was apparent during the tests and in the load/displacement result
plots.
2.2 Phenolic resin tension test
Tensile tests were performed on resol phenolic resin plates (Hirenol® KRD-HM2, Kolon Industries
Inc., South Korea). These tests were performed to characterize a constituent material present in the
composition of Nomex honeycomb cores. Phenolic resin was mixed with 25 wt% ethanol, and 250
mm × 185 mm × 2 mm thick plates were fabricated. To prevent excessive foaming, a gradual cure
cycle was used as follows: 4 hours each at 50°C, 60°C, 65°C, 70°C, 75°C and 80°C, 2 hours each at
90°C and 100°C, 1 hour at 120°C, and 1.5 hours at 150°C [25]. Shrinkage of approximately 2.5% was
observed upon curing. Five tensile specimens were cut according to the ASTM D638-03 standard
with an overall length of 200 mm and a narrow section width of 13 mm [26]. Both specimen main
surfaces were polished up to 3 μm grit; the average final specimen thickness was 1.76 mm.
Glass/epoxy laminates were used as tabs, measuring 42 mm × 19 mm × 1.6 mm. Specimens were
installed with a 2-element 90° cross rosette strain gage (model UFCA-3-17-3L, Tokyo Sokki Kenkyujo
Co. Ltd., Tokyo, Japan) at mid-length after sanding the surface with #1000 sandpaper. A constant
head displacement rate of 5 mm/min was applied in the tests. Strain gage signals were recorded at
Manuscript preprint
10 Hz through a NEC AS1203 strain amplifier (NEC Corporation, Tokyo, Japan) and a National
Instrument USB-6251 data acquisition module (National Instruments Corporation, Austin, USA), as
was the case with all strain gages used in this paper. During the tensile tests, the specimens failed
at the grips; therefore, the measured strength was not considered. The cured resin was brittle and
it could crack while cutting specimens too quickly. Nonetheless, a sufficiently high load was reached
in the tests to measure significant modulus and Poisson’s ratio values.
2.3 Nomex/phenolic tension test
Specimens made of Nomex paper dipped into phenolic resin were fabricated. Specimens were
fabricated with the paper in the roll direction (0°), transverse direction (90°), and 45°. Nomex paper
specimen sections initially measured 50 mm (width) × 350 mm (length) × 0.05 mm (thickness).
Paper sections were secured straight in a metal holder (Figure 3a). This holder was then dipped in
a bath of phenolic resin (same resin as in section 2.2) mixed with 25 wt% ethanol for 5 minutes
(Figure 3b). After removal from the bath, the holder was left to drip above the bath for 5 minutes.
The specimens were then cured in an oven while still in the holder. The cure cycle was as follows: 2
hours at room temperature, 1 hour each at 70°C, 100°C and 120°C, and finally 1.5 hours at 150°C
[25]. Initially, seven specimens were installed at once in the holder, but later only three specimens
were installed to prevent accidental contact between the dipped paper sections. We therefore advise
in this case to use a minimum gap of approximately 8 mm between paper sections. After removal
from the holder, the lower 10 mm and longitudinal extremities of the specimens were cut-out, giving
a final specimen area of 40 mm (width) × 290 mm (length). Tabs consisting of 25 mm × 40 mm
#100 sandpaper sections adhered with epoxy were used, giving an effective gage length of 240 mm.
From random measurements, we observed that the specimen thickness varied by approximately
0.02 mm from top to bottom widthwise. This may be due to a gravity effect during the d ipping
process. Specimens were installed with a 2-element 90° cross rosette strain gage at mid-length
(model FCA-3-11-1L, Tokyo Sokki Kenkyujo Co. Ltd., Tokyo, Japan). Three specimens for each paper
direction were tested with a constant head displacement rate of 1 mm/min.
2.4 Nomex honeycomb core test
Flat sandwich panel specimens with Nomex honeycomb cores were prepared to perform flatwise
tension and compression tests. In the tests, strain gages were installed on the honeycomb walls.
These tests served to provide additional core material stress/strain behavior data and also allowed
for the investigation of possible differences in the material’s tensile and compressive moduli. With
aramid fibers for example, there is evidence that contrary to strain stiffening in tension, strain
softening can occur in compression [27]. The sandwich panels tested had faces consisting of
glass/epoxy 3/1 weave prepreg fabric laminates (GEP 224, SK Chemicals, South Korea). These
laminates were first vacuum bag-cured alone in an autoclave at 3.0 atm with a stacking sequence
of [0/90]2S and total thickness of 1.0 mm. Two Nomex honeycomb cores were evaluated: both with
a 4.8 mm cell size and 25.4 mm height and with densities of 32 kg/m3 and 64 kg/m3 (models HD322
and HD342, M.C. Gill Corporation, El Monte, USA). These cores were provided by Korea Aerospace
Manuscript preprint
Industries Ltd. The faces were lightly sanded (#1000 grit sandpaper) and adhered to the cores with
an epoxy adhesive film (Hysol EA9696.030, Henkel Corporation, Düsseldorf, Germany) in vacuum
bag autoclave molding with a maximum pressure of 2.0 atm. The bag vacuum was released when
the autoclave pressure climbed past 1.0 atm. The faces were adhered to the core one at a time so
that the effect of gravity on the adhesive line would be the same on both sides. The specimen’s
nominal area dimensions were 50 mm × 50 mm with a total height of 27.4 mm. To first perform a
flatwise tensile test, specimens were adhered to steel blocks with a liquid epoxy adhesive cured at
room temperature. Two small strain gages were then installed on the specimens (model UFLK-1-
17-3L, Tokyo Sokki Kenkyujo Co. Ltd., Tokyo, Japan), one on each side parallel to the L-direction, on
a honeycomb double-wall section (Figure 5). Flatwise tensile tests were performed at a machine
head displacement rate of 0.2 mm/min and stopped before failure when the recorded gage strain
reached approximately 0.3%. Specimens were then removed from the flatwise tensile test jigs,
keeping the adhered steel block in place. Flatwise compression tests were then performed on these
specimens until failure with the same machine head displacement rate of 0.2 mm/min [28]. Three
specimens of each core type were tested.
2.5 Density measurement
The density of the specimens tested was measured by taking their weight in air and in water with
an electronic density meter (VIBRA DME, Shinko Denshi Co., Jakarta, Indonesia). This density meter
model automatically calculates the density from these two weight measurements. The density of
the Nomex/phenolic specimen was measured after the tensile test from an un-cracked area section
of 40 mm × 50 mm cut from each specimen. The average thickness (t) of this section was then
calculated with the following formula: t = m/(ρA), where m is the specimen’s dry mass, ρ its
measured density, and A its surface area. The average constituent material density mat) of five
different Nomex honeycomb cores was also measured. This measurement was done by submerging
a honeycomb core section in the density meter. Core specimens with an area measuring
approximately 6 cm × 6 cm were used. When the core specimen was submerged in the density
meter, it was slightly shaken to dislodge any air bubbles from it.
3. Test results and discussion
3.1 Nomex paper
Figure 7 presents representative stress/strain curves from Nomex paper tensile tests. To calculate
the stress, the recorded load was divided by the specimen’s section, considering an actual thickness
of 0.056 mm, based on measurement data from the manufacturer [24]. Strain was calculated by
taking the machine displacement divided by the specimen’s initial gage length of 300 mm. The
plastic-like behavior observed in the curves is similar to what Foo et al. previously reported [23].
Likewise, wrinkling of the paper began at approximately the mid-level force amplitude. Every
specimen we tested failed in its middle length portion, away from the tabs. The tensile modulus
Manuscript preprint
was calculated from a linear regression in the 0.20.4% strain interval. Table 1 presents the average
test results for the different paper directions. Compared to Foo et al. [23], the obtained average
moduli test results differ by 723% and reveal a higher orthotropic ratio (1.62 vs 1.38 for Foo et al.)
(Table 2). While it is premature to make a conclusion because of limited data, these differences
could possibly be caused by the different paper thickness tested (0.05 mm vs 0.13 mm for Foo et
al.). In terms of strength, both results are comparable to published data from the manufacturer [24].
The listed strength values of Foo et al. were determined by estimating the maximum forces on the
test curves in their published work.
3.2 Phenolic resin
The phenolic resin tensile test results displayed a linear elastic stress/strain behavior, as expected.
The tensile modulus and Poisson’s ratio were calculated from a linear regression of the stress/strain
and εyx slope, respectively, in the 0.1%-0.3% strain range. The average tensile modulus was 4.94
GPa (Sn-1 = 0.20 GPa, C.V. = 4.05%), and the average Poisson’s ratio was 0.389 (Sn-1 = 0.005, C.V. =
1.35%), measured from five specimens. The modulus obtained is higher than a stated guideline
value of 3.9 GPa [29] but slightly lower than the measured result of Redjel [30] (5.16 GPa). Giglio et
al. [14] recently used a modulus value of 3 GPa in their numerical model. The Poisson ratio result
of ν = 0.389 compares to a measured result of 0.36 from Redjel [30]. By isotropic material theory, a
shear modulus of 1.78 GPa was calculated from the current results {G = E/(1+2ν)}. The average
cured resin density, measured from three specimens of approximately 13 mm × 40 mm × 1.76 mm
dimensions, was ρphenol = 1.342 g/cc.
3.3 Nomex/phenolic
Representative stress/strain curves from the Nomex/phenolic tension tests are presented in Figure
8. Linear elastic behavior was observed until failure. Compared to plain Nomex paper (Figure 7), the
addition of a phenolic resin coating eliminated plastic deformation and greatly reduced maximum
strain. This behavior was also observed by Hähnel [31], who reported semi-qualitative stress/strain
slopes of Nomex paper impregnated with phenolic resin. When Hähnel used only a single resin
impregnation, the material retained some plastic behavior, but with two resin impregnations, the
material exhibited completely linear elastic behavior. Considering this result, we assumed that our
single resin impregnation was relatively thick in comparison. For the 0° Nomex/phenolic specimens,
the average failure strength of these specimens (42.9 MPa) corresponds to a force of 204.5 N. Let
us assume that, in this test, the phenolic coating is the first to fail. This assumption is based on the
Nomex/phenolic specimen’s relatively low failure strain compared to plain Nomex paper (0.83%
vs
7.27%). Failure of the coating would then transfer a total stress of 89.6 MPa {204.5 N / (40.75 ×
0.056) mm2} to the paper alone, higher than its previously measured average strength of 64.2 MPa.
In this scenario, the phenolic coating and the paper would fail virtually simultaneously. This scenario
also assumes a simplified representation where there is no phenolic resin impregnation in the paper.
The tensile modulus and Poisson’s ratio were calculated from a linear regression of the stress/strain
Manuscript preprint
and εyx slopes, respectively, in the 0.2%-0.4% strain range. From the density measurement,
calculated average individual specimen thicknesses were in the range of 0.1090.145 mm. These
thicknesses were in agreement with random individual caliper measurements on the specimens.
Average test results for the different paper directions are presented in Table 3. The moduli and
Poisson’s ratios here listed are raw test data. Later in this text, the local effect of the strain gage will
be modeled in FEM analysis. Varying specimen thickness will also be normalized in the model. The
test data will therefore be corrected to obtain typical Nomex honeycomb core constituent material
properties.
3.4 Nomex honeycomb core constituent material density
The measured constituent material densities mat) of five Nomex honeycomb cores are listed in
Table 4. Despite the different core densities and manufacturers evaluated, constituent material
densities are fairly close, within a range of 1.111.16 g/cc. Using the core and material densities, the
cell size (H), and a theoretical perfect hexagon cell geometry, a theoretical honeycomb constant
single-wall thickness (tth) can be calculated (equation 1). This is an idealized representation because
resin accumulation in cell corners is not considered. From these calculations, a slight trend is
apparent: the higher the wall thickness, the higher the material density. Only the 'McGill HD322'
specimen departs from this trend. This trend is consistent with the fact that the density of Nomex
paper increases with thickness: for example, for t = 0.056 mm, the density is ρ = 0.72 g/cc, and for
t = 0.079 mm, the density is ρ = 0.80 g/cc [24]. Furthermore, this trend is consistent if the increase
in wall thickness comes from extra phenolic resin coating because its density is greater than Nomex
paper (ρphenol = 1.342 g/cc).

 
 (1)
3.5 Nomex honeycomb core flatwise test
Figure 9 shows the global core stress/strain behavior in the flatwise tension and compression tests.
Core stress was taken as the applied load divided by the core area (A = 50 mm × 50 mm), and core
strain was taken as the machine displacement divided by the core height (T = 25 mm). In
compression, the 32 kg/m3 core yielded at approximately 85% (-0.6 MPa) of its maximum strength,
while the 64 kg/m3 core had a more brittle behavior, with no significant yield before failure.
Combined stress/strain curves derived from the strain gage signals are presented in Figure 10.
Honeycomb wall stress was considered in this case and was calculated by taking the applied load
divided by an idealized honeycomb wall section area [20]. Conversion from core stress to wall stress
was calculated as in equation (2), with l = 2.75 mm the cell wall length, φ = 60° the hexagonal cell
angle, and wall thicknesses of t(32kg/m³) = 0.0489 mm and t(64 kg/m³) = 0.0924 mm (from Table 4).

   
 {36.54 (32 kg/m3), 19.33 (64 kg/m3)} (2)
Manuscript preprint
On the tensile side of the strain gage signals, linear elastic behavior was strictly observed for the
32 kg/m3 core, while the 64 kg/m3 core had one or two strain gages showing some nonlinear drift.
We attribute this behavior to experimental imprecision, namely load misalignment and strain gage
location imprecision on the honeycomb wall. In compression, at low stress, behavior is generally
linear, while some strain gages showed pronounced nonlinear drift with increasing stress. Again, we
attribute this drift to experimental imprecisions. We found that our flatwise compression test was
sensitive to specimen alignment (placing the specimen center in-line with the load). Keeping the
steel blocks in the test, and hence adding height, might have amplified this sensitivity. For the 32
kg/m3 core, even the strain gage signals that remained fairly linear with increasing stress diverged
at approximately 85% of the maximum compressive stress (around -0.6 MPa). This behavior is
consistent with the yielding observed in the global core stress/strain curves. The 64 kg/m3 core’s
strain gage signals remained fairly linear until failure; this behavior is again consistent with the
global core brittle behavior observed. A linear regression of the stress/strain slope in the 0.1%-0.2%
strain range (tension) and {-0.1%,-0.2%} strain range (compression) was calculated from the strain
gage signals. We define this quantity as the cell wall elastic modulus, and individual results are
presented in Table 5. The 32 kg/m3 core had similar tension/compression average moduli (Etens =
5.29 GPa, Ecomp = 5.30 GPa), while the 64 kg/m3 core was 7.22% softer in compression (Etens = 5.26
GPa, Ecomp = 4.88 GPa). In both cases, the variability on the moduli was relatively high (C.V. 10%),
and few specimens were tested (3 each). We therefore will not make any definite conclusions from
the differences in tension and compression moduli. As for the Nomex/phenolic test results, the
flatwise tests will be modeled in FEM and the results compensated for the strain gage local stiffening
effect. Because of the Nomex honeycomb core’s expansion fabrication process [32], it is most likely
that the Nomex paper’s 90° direction will be aligned with the core’s through-thickness direction.
This direction was the main direction of the load and strain measurement in these core flatwise
tests. It will thus be possible to compare identified 90° direction properties from the Nomex/phenolic
specimens and the core flatwise tests.
4. Modeling
4.1 Nomex/phenolic tension test model
A finite element model of the Nomex/phenolic tension tests was built in order to compensate for
the local stiffening effect expected in the strain gage area. To achieve this, the strain gage material
was included in the model. Based on the properties of epoxy films [33], the following parameters
were used to model the strain gage: E = 2.5 GPa, ν = 0.36, and a 0.03 mm thickness. For every
tensile test group (i.e., 0°, 45°, and 90°), a model was built that considers each group’s particular
average thickness (from Table 3) to determine the extra phenolic coating thickness in the model
(Figure 12). The models therefore also served to normalize the phenolic coating thickness of the
different specimen groups. From microscope observation of the Hexcel 80 kg/m3 core, we
determined the average total wall thicknesses to be approximately 0.08 mm, of which approximately
0.055 mm is the Nomex paper only. The properties of the 0.08 mm thick center portion (material
Manuscript preprint
to identify) were varied by trial and error in order to match the model’s effective properties at the
strain gage to the test results. We note that this 0.08 mm thick material includes any resin
impregnation in the Nomex paper. The test moduli could be matched by adjusting almost
independently one material property in the model: E1 for E, E2 for E90°, and G12 for E45°. The
simulated effective gage Poisson’s ratios could not be matched exactly to all experimental results
simultaneously. The model’s Poisson’s ratio was therefore chosen based on a least mean square
calculation of the model’s deviation from the Poisson’s ratio of the three test results (0°, 45°, and
90°). Identified Nomex core constituent material properties are listed in Table 6. The E2 modulus
that was identified (3.22 GPa) is lower than the one obtained by Staal [4] (3.81 GPa) and the 3.5
GPa value used by Asprone et al. [13] and Roy et al. [19] in their models. The shear modulus that
was obtained (1.26 GPa) is also lower than what was previously identified in Roy et al. (1.68 GPa)
[19]. These moduli differences will be discussed further in section 4.3. The in-plane Poisson’s ratio
that was identified (0.24) differs from the value of 0.4 frequently used in Nomex core FEM model
simulations [6-8, 18]. However, the in-plane Poisson’s ratio is closer to the value of 0.2 previously
identified by Roy et al. [19]. In the absence of any stated rationale, the Poisson`s ratio value of 0.4
previously cited appears taken from phenolic resin property, or from earlier work (e.g., [12]). Given
the Poisson’s ratio of the phenolic resin we used (0.389), the current Nomex/phenolic results (ν =
0.24) would imply that the Nomex paper itself has a relatively lower Poisson’s ratio value. For
verification, let’s assume the previously described 0.08 mm thick Nomex/phenolic material as layered.
This material would have a 0.055 mm thick Nomex paper core with 0.0125 mm thick phenolic resin
coatings and no resin impregnation in the Nomex paper. From classical laminate theory calculations
[34], we found that a Nomex paper Poisson’s ratio value of 0.193 is required to obtain a
Nomex/phenolic Poisson’s ratio of 0.24. Under these assumptions, this 0.193 value is comparable or
at the low range of reported Poisson’s ratio values for wood-based paper materials [35-36].
4.2 Core flatwise test model
The core flatwise tests were modeled in FEM to calibrate the strain gage measurement results. A
meso-scale model was built with shell elements representing the honeycomb cell geometry. This
type of model can consider several aspects of an actual core’s variable nature, such as cell wall
thickness variability [13, 37-38], cell wall curvature (in-plane, out-of-plane) [4, 20-21, 39], material
property variability within a cell wall [13, 40], resin accumulation at the cell corners [14], and other
random geometrical imperfections [15]. These characteristics can vary depending on the core
manufacturer. To begin with a relatively simplified representation, we considered a constant cell wall
thickness and some in-plane wall curvature at the cell corners (fillets), similar to Staal [4]. Staal
showed that using filleted cell corners can have a significant effect on the model’s in-plane face
wrinkling buckling load. From microscope observations of the cores, a cell fillet ratio (ψ) of 0.2 was
implemented in the model geometry, defined as ψ = (l-L)/2l, with l the idealized hexagonal cell wall
length and L the actual straight portion of the cell wall (Figure 11). Linear 4-node shell elements
were used, with a mesh division of 9 elements (in-plane wise) by 106 elements (height wise) for
each cell wall segment. This mesh resulted in an element size of approximately 0.2 mm × 0.24 mm,
Manuscript preprint
which was considered appropriate based on published mesh convergence analysis of similar models
[15]. The core’s bottom node displacements (ui) were locked in the X, Y, and Z directions. The core’s
top node displacements were coupled in the X, Y, and Z directions. With this we considered that
the in-plane stiffness of the core was negligible compared to the faces. This coupling permitted a
single vertical force to be applied on a node at the geometrical center at the top of the core. The
particular strain gage used was also modeled assuming polyimide film properties (E = 5.32 GPa, ν
= 0.35, from [41]) and a thickness of 0.05 mm. For both cores tested (32 kg/m3 and 64 kg/m3), the
thicknesses of single-wall elements closest to cell corners (t3) and second closest to cell corners (t4)
were adjusted to obtain the exact core density desired in the model. This adjustment is also
designed to represent resin accumulation present in that area. The cell wall thicknesses were t1 =
0.0489 mm, t2 = 2·t1, t3 = t2, and t4 = 0.05632 mm for the 32 kg/m3 core, and t1 = 0.0924 mm, t2 =
2·t1, t3 = t2, and t4 = 0.12395 mm for the 64 kg/m3 core. The Nomex/phenolic properties obtained
in section 4.1 were used as a starting point for the core cell wall material properties. The strain
result of the strain gage elements was then averaged and compared to experimental results. To
obtain an exact match to the strain gage experimental results, moduli were all scaled by a constant.
For the 32 kg/m3 core, a proportion of 1.293 applied to the moduli (E1 = 6.73 GPa, E2 = 4.16 GPa,
G12 = 1.63 GPa) replicated the exact modulus measured at the gage; for the 64 kg/m3 core, this
proportion was 1.482 (E1 = 7.71 GPa, E2 = 4.77 GPa, G12 = 1.87 GPa). Compared to the 90° direction
modulus identified in section 4.1 (E2 = 3.22 GPa), the moduli identified with this approach are 29.2%
(32 kg/m3 core) and 48.1% (64 kg/m3 core) higher. Our best hypothesis for this difference is that
the cell wall material of the cores tested simply has greater stiffness than the specimens we
fabricated. Nomex paper can be tailored and optimized for a given application. This tailoring can
involve adjusting the proportion of fibers, pulp, and binder in the paper, as well as the percentage
of voids. Such tailoring has been shown to have a significant effect for honeycomb cores [42].
Moreover, Nomex paper type 412 is typically used in honeycomb cores [13, 15, 25]. We used general
usage Nomex paper type 410 to prepare our specimens, like other authors [43-44], because this
type was more easily available to us. This different paper grade and the fact that we applied one
thick phenolic coating at once in our specimens (e.g. [45]) may be the reason that our specimens
are softer than the core cell wall material. It is also worth noting that the McGill cores tested are
among the stiffest comparable Nomex honeycomb cores commercially available [32, 46].
4.3 Representative Nomex core model
The material properties identified in section 4.1 were also evaluated for a core with a cell wall
thickness comparable to the one used in the identification process (0.08 mm). The Hexcel 80 kg/m3
density, 3.18 mm cell size core was chosen as a basis because we previously determined that its cell
has the same nominal cell wall thickness of 0.08 mm. A meso-scale shell FEM model of the core
was built in the same fashion as described in the previous section. A cell fillet ratio (ψ) of 0.17 was
implemented based on microscope observations. The cell hexagon angle was also measured and
found to be φ = 47.3° with an actual cell size (H) of 3.38 mm. The thicknesses of the single-wall
elements closest to cell corners were again adjusted in order to match the measured core density
Manuscript preprint
(76.6 kg/m3 from Table 4). This model served to compare simulated core moduli with test data
published by the manufacturer. The model’s simulated core moduli were determined by applying a
single force (Fi) on a top node at the geometric center and measuring the corresponding core
deformation. For compression, the model had a square area of 50 mm × 50 mm, typical of a flatwise
test. In the case of shear, it is recommended that the specimen length be 1012 times that of its
thickness [47-48]. Therefore in shear loading, the model had a rectangular area of 50 mm × 138
mm. The simulated core moduli are presented in Table 7, case 2, and they are all lower than the
manufacturer’s test data. As stated before, our best hypothesis for this difference is that the
specimens we prepared may have a lower stiffness than commercial honeycomb materials. The
material moduli in the model were subsequently scaled to get closer to the manufacturer’s core
moduli data. The E2 modulus was scaled to exactly match the core’s compression modulus. Scaling
the E1 modulus had little influence on the different simulated core moduli, so it was left unchanged.
The G12 modulus was scaled to approach the core’s shear moduli, as it was not possible to
simultaneously obtain a perfect match to both core main direction shear modulus. An optimal G12
value was therefore determined from a least mean square calculation of the percentage deviation
from the manufacturer’s data. A much higher shear modulus was required (G12 = 1.85 GPa), and we
observed that, compared to the manufacturer’s test data, the model is relatively stiffer in the L-
direction than in the W-direction (Table 7, case 3). We verified that the rectangular area size chosen
showed convergence within below 1% on the G12 value identified, compared to a slightly smaller
size model. It was also verified that shear loading on the square model would generate a 6%
increase difference in the G12 value identified. Based on the diverging directional core shear stiffness
obtained, we reasoned that it is more likely that the manufacturer data comes from a core with
nominal dimensions, that is: φ = 60°, H= 3.18 mm, ρcore = 80 kg/m3, and ψ = 0.2 (from [4]). Therefore,
another model was built with these parameters. With an increased cell hexagon angle (φ = 60°), we
could expect a correction in the GL/GW ratio according to analytical models [49]. This was the case
as shown in Table 7, case 4, where the shear moduli are now within 3.3% of the test data. Calibrated
E2 and G12 values are also slightly lower with this model, most likely in great part due to the
increased core density modeled. Another possible consideration in the model is the double-wall (t2)
thickness allocation. In reality, this double-wall may not have exactly double the thickness of a
single-wall (t1). The double-wall should have twice the Nomex paper thickness as a single-wall, but
sensibly the same phenolic coating thickness. Microscope observations of core cells confirms this
reasoning, and on average we measured a thickness of t2 = 1.77·t1. This cell wall thickness proportion
was incorporated into the nominal dimensions model (φ = 60°), which was also adjusted to obtain
a 80 kg/m3 core density (t1 = 0.08 mm, t2 = 0.14175 mm, t3 = t2, t4 = 0.11368 mm). The simulated
shear moduli ratio obtained with this configuration were then within 0.6% of the target value (Table
7, case 5). For the sake of obtaining a perfect match, a thickness proportion of t2 = 1.74·t1 was used
(t1 = 0.08 mm, t2 = 0.0001392 mm, t3 = t2, t4 = 0.0001206 mm). This configuration gave simulated
core moduli that exactly matched the manufacturer’s test data, along with corresponding calibrated
cell wall material elastic properties (Table 7, case 6). This result demonstrated possible approaches
to model calibration and also the extent of their effect, specifically in terms of shear moduli ratio.
Manuscript preprint
Using slightly different material properties for the double-wall may also be another way to calibrate
the model, ideally with some analytical rationale or test data to back it. The obtained calibrated cell
wall material elastic properties are comparable to some used in recent published work on meso-
scale FEM modeling of Nomex honeycomb cores: E2 = 3.60 GPa (3.50 GPa in [13], 3.52 GPa in [19]),
G12 = 1.76 GPa (1.68 GPa in [19]), and ν12 = 0.24 (0.2 in [19]).
5. Conclusion
In this work, tensile tests were first performed on Nomex paper, phenolic resin, and Nomex paper
coated with phenolic resin. Nomex paper test results were comparable to previously published test
data. These results also confirmed that paper properties are dependent on the paper’s nominal
thickness, at least in terms of strength. Measured phenolic resin elastic properties were in the range
of published results. Phenolic resin coated Nomex paper displayed a strictly linear elastic behavior;
we attribute this behavior to the relatively thick resin coating that was used. These Nomex/phenolic
test results were calibrated in finite element modeling (FEM) to account for strain gage local
stiffening effects and for specimen thickness normalization. The identified Nomex/phenolic elastic
properties were E1 = 5.20 GPa, E2 = 3.22 GPa, G12 = 1.26 GPa, and ν12 = 0.24. Flatwise tension and
compression tests were also performed on Nomex honeycomb cores with strain gages installed.
Finite element models of these tests were also built and E2 material values of 4.16 GPa or 4.77 GPa
were determined from the models. We attribute these relatively stiffer values to a more optimized
material formulation in the commercial Nomex cores compared to our fabricated Nomex/phenolic
test specimens. The identified Nomex/phenolic material property set was then used in a finite
element model of a core. Simulated core moduli were compared to manufacturer’s test data. Scaling
of the material moduli E2 and G12 was required to approach the test data. It was also observed that
the honeycomb cell’s double-walls do not exactly have double the thickness of its single-walls. This
observation was also implemented in the model, and it was possible to achieve a perfect match
between core simulations and test results. The calibrated cell wall material elastic properties in this
matching model were E1 = 5.20 GPa, E2 = 3.60 GPa, G12 = 1.76 GPa, and ν12 = 0.24. We consider
that some level of model calibration should be expected with meso-scale Nomex honeycomb core
modeling. The need for calibration may come from the level of modeling sophistication used
(imperfections, cell corners), core construction differences among manufacturers, or core test results
variability.
Acknowledgment
This work was supported by the Priority Research Centers Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2009-
0094104). This research was financially supported by the Ministry Of Trade, Industry & Energy
(MOTIE), Korea Institute for Advancement of Technology (KIAT) and Dong-Nam Institute For Regional
Program Evaluation (IRPE) through the Leading Industry Development for Economic Region. Part of
Manuscript preprint
this work was presented at the 17th International Conference on Composite Structures (ICCS-17),
17-21 June 2013, in Porto, Portugal.
References
1. ASTM Standard C297. Standard test method for flatwise tensile strength of sandwich constructions.
ASTM International. West Conshohocken, PA; 2004.
2. Khan S, Loken HY. Bonding of sandwich structures - The facesheet/honeycomb interface - A
phenomenological study. In: Proceedings of SAMPE 2007, Baltimore, 3-5 June 2007, pp. 1-9.
3. HexWebTM Honeycomb Sandwich Design Technology. Hexcel Composites. Duxford, UK; 2000.
4. Staal RA. Failure of sandwich honeycomb panels in bending. PhD Thesis, University of Auckland,
New Zealand; 2006.
5. Menna C, Zinno A, Asprone D, Prota A. Numerical assessment of the impact behavior of
honeycomb sandwich structures. Compos Struct 2013;106:326-39.
6. Manes A, Gilioli A, Sbarufatti C, Giglio M. Experimental and numerical investigations of low velocity
impact on sandwich panels. Compos Struct 2013;99:8-18.
7. Foo CC, Chai GB, Seah LK. A model to predict low-velocity impact response and damage in
sandwich composites. Compos Sci Technol 2008;68:1348-56.
8. Giglio M, Gilioli A, Manes A. Numerical investigation of a three point bending test on sandwich
panels with aluminum skins and Nomex™ honeycomb core. Comput Mater Sci 2012;56:69-78.
9. Song KI, Choi JY, Kweon JH, Choi JH, Kim KS. An Experimental Study of the Insert Joint Strength
of Composite Sandwich Structures. Compos Struct 2008;86:107-13.
10. Heimbs S, Pein M. Failure behaviour of honeycomb sandwich corner joints and inserts. Compos
Struct 2009;89:575-88.
11. Park YB, Kweon JH, Choi JH. Failure characteristics of carbon/BMI-Nomex sandwich joints in
various hygrothermal conditions. Compos: Part B 2014;60:213-21.
12. Zang J, Ashby MF. The out-of-plane properties of honeycombs. Int J Mech Sci 1992;34:475-89.
13. Asprone D, Auricchio F, Menna C, Morganti S, Prota A, Reali A. Statistical finite element analysis
of the buckling behavior of honeycomb structures. Compos Struct 2013;105:240-55.
14. Giglio M, Manes A, Gilioli A. Investigations on sandwich core properties through an
experimental-numerical approach. Compos: Part B 2012;43:361-74.
15. Heimbs S. Virtual testing of sandwich core structures using dynamic finite element simulations.
Comput Mater Sci 2009;45:205-16.
16. Aktay L, Johnson AF, Kröplin BH. Numerical modelling of honeycomb core crush behavior. Eng
Fract Mech 2008;75:2616-30.
17. Kaman MO, Solmaz MY, Turan K. Experimental and Numerical Analysis of Critical Buckling Load
of Honeycomb Sandwich Panels. J Compos Mater 2010;44:2819-31.
18. Aminanda Y, Castanie B, Barrau JJ, Thevenet P. Experimental analysis and modeling of the
crushing of honeycomb cores. Appl Compos Mater 2005;12:213-27.
19. Roy R, Nguyen KH, Park YB, Kweon JH, Choi JH. Testing and modeling of Nomex™ honeycomb
Manuscript preprint
sandwich Panels with bolt insert. Compos: Part B 2014;56:762-69.
20. Roy R, Kweon JH, Choi JH. Meso-scale finite element modeling of Nomex™ honeycomb cores.
Adv Compos Mater 2013;23:17-29.
21. Kress G, Winkler M. Honeycomb sandwich residual stress deformation pattern. Compos Struct
2009;89:294-302.
22. Tsujii Y, Tanaka K, Nishida Y. Analysis of mechanical properties of aramid honeycomb core. Trans
Jpn Soc Mech Eng 1995;61:1608-14.
23. Foo CC, Chai GB, Seah LK. Mechanical properties of Nomex material and Nomex honeycomb
structure. Compos Struct 2007;80:588-94.
24. Nomex® type 410 technical data sheet. E.I. Du Pont de Nemours, USA, 2012.
25. Kim KS, Chin IJ, Sung IK, Min KS. Curing of Nomex-phenolic and kraft-phenolic honeycombs.
Kor Polym J 1995;3:35-40.
26. ASTM Standard D638-03. Standard test method for tensile properties of plastics. ASTM
International. West Conshohocken, PA; 2003.
27. Andrews MC, Lu D, Young RJ. Compressive properties of aramid fibres. Polym 1997;38:2379-88.
28. ASTM Standard C365-03. Standard test method for flatwise compressive properties of sandwich
cores. ASTM International. West Conshohocken, PA; 2003.
29. Dieter GE. ASM Handbook Volume 20: Materials selection and design. ASM International, 1997.
30. Redjel B. Mechanical properties and fracture toughness of phenolic resin. Plast Rub Compos
Process Appl 1995;24:221-28.
31. Hähnel F, Wolf K. Evaluation of the material properties of resin-impregnated Nomex® paper as
basis for the simulation of the impact behaviour of honeycomb sandwich. In: Book of abstract of
CompTest2006, Porto, 10-12 April 2006, pp. 168-9.
32. HexWeb® honeycomb attributes and properties. Hexcel Corporation. Stamford, CT; 1999.
33. DaSilva LFM, Adams RD. Measurement of the mechanical properties of structural adhesives in
tension and shear over a wide range of temperatures. J Adhes Sci Technol 2005;19:109-41.
34. Berthelot JM. Mécanique des matériaux et structures composites. Le Mans, France: ISMANS,
2010.
35. Yokoyama T, Nakai K. Evaluation of in-plane orthotropic elastic constants of paper and
paperboard. In: Proceeding of 2007 SEM Annual Conference & Exposition on Experimental and
Applied Mechanics, Springfield, 3-6 June 2007, pp. 1505-11.
36. Szewczyk W. Determination of Poisson’s Ratio in the Plane of the Paper. Fibres Text East Eur
2008;16:117-20.
37. Yang MY, Huang JS. Numerical analysis of the stiffness and strength of regular hexagonal
honeycombs with plateau borders. Compos Struct 2004;64:107-14.
38. Karakoç A, Santaoja K, Freund J. Simulation experiments on the effective in-plane compliance of
the honeycomb materials. Compos Struct 2013;96:312-20.
39. Yang MY, Huang JS, Hu JW. Elastic buckling of hexagonal honeycombs with dual imperfections.
Compos Struct 2008;82:326-35.
Manuscript preprint
40. Fischer S, Drechsler K, Kilchert S, Johnson A. Mechanical tests for foldcore base material
properties. Compos: Part A 2009;40:1941-52.
41. Djakaria R, Chandran BI, Gordon MH, Schmidt WF. Determination of Young’s modulus of thin
films used in embeded passive devices. In: Proceedings of the 47th Electronic Components and
Technology Conference, San Jose, 18-21 May 1997, pp. 745-49.
42. Nomoto K. Aramid honeycombs and a method for producing the same. European patent
specification (EP 1 152 084 B1), 2001.
43. Shafizadeh JE, Seferis JC. Scaling of honeycomb compressive yield stresses. Compos: Part A
2000;31:681-88.
44. Wu YJ, Seferis JC, Lorentz V. Evaluations of an Aramid Fiber in Nonwoven Processes for
Honeycomb Applications. J Appl Polym Sci 2002;86:1149-56.
45. Shafizadeh JE, Seferis JC. Characterization of phenolic resins for composite honeycomb
applications. In: Proceeding of ICCM 12, Paris, 5-9 July 1999, paper 1285.
46. Gillcore™ HD Honeycomb. Product data sheet. M.C. Gill Corporation. El Monte, CA; 2013.
47. Hong ST, Pan J, Tyan T, Prasad P. Quasi-static crush behavior of aluminum honeycomb specimens
under non-proportional compression-dominant combined loads. Int J Plast 2006;22:1062-88.
48. ASTM Standard C273-00. Standard test method for shear properties of sandwich core materials.
ASTM International. West Conshohocken, PA; 2000.
49. Hohe J, Becker W. Effective stress-strain relations for two-dimensional cellular sandwich cores:
Homogenization, material models, and properties. Appl Mech Rev 2002;55:61-87.
Figures
(a)
Manuscript preprint
(b)
Figure 1: (a) Nomex paper tension test set-up, (b) Nomex paper roll.
Figure 2: Phenolic resin tension test specimens and tension test set-up.
Manuscript preprint
(a)
(b)
Figure 3: (a) Nomex paper sections in holder; (b) Nomex paper sections in phenolic resin
bath.
Figure 4: Nomex/phenolic tensile test set-up.
Manuscript preprint
(a)
(b)
Figure 5: (a) Flatwise tension test; (b) Flatwise compression test.
Manuscript preprint
Figure 6: Electronic density meter.
Figure 7: Nomex T410 paper representative stress/strain curves.
Manuscript preprint
Figure 8: Nomex/phenolic specimen representative stress/strain curves.
(a)
Manuscript preprint
(b)
Figure 9: Flatwise tension and compression core stress/strain curves: (a) 32 kg/m3 core; (b)
64 kg/m3 core.
(a)
Manuscript preprint
(b)
Figure 10: Flatwise tension and compression strain gage data: (a) 32 kg/m3 core; (b) 64
kg/m3 core.
Figure 11: Honeycomb core cell geometry description.
Manuscript preprint
Figure 12: Representation of the Nomex/phenolic tension test FEA model.
Figure 13: Nomex honeycomb core FEA model (compression model dimensions pictured).
Manuscript preprint
Tables
Table 1: Nomex T410 paper tension test average results.
Paper direction
E [GPa]
σu [MPa]
εu [%]
Mean
3.18
64.2
7.27
Sn-1
0.088
2.6
0.75
C.V.
2.85%
4.04%
10.3%
Mean
2.36
42.6
6.50
45°
Sn-1
0.10
0.53
0.57
C.V.
4.24%
1.23%
8.71%
Mean
1.96
33.6
6.69
90°
Sn-1
0.069
0.61
0.85
C.V.
3.52%
1.82%
12.8%
Table 2: Nomex paper tensile property comparison.
Paper nominal thickness
Source
E0 [GPa]
E90 [GPa]
σ0 [MPa]
σ90 [MPa]
0.05 mm
Current results
3.18
1.96
64.2
33.6
Dupont [23]
69.8
32.2
0.13 mm
Foo et al. [22]
3.40
2.46
108.5
54.2
Dupont [23]
103.7
50.0
Table 3: Nomex/phenolic tension test raw results.
Paper
direction
E [GPa]
ν
σu [MPa]
εu [%]
ρ [g/cc]
t [mm]
Mean
5.31
0.292
42.9
0.829
1.334
0.117
Sn-1
0.258
0.060
4.0
0.099
0.038
0.009
C.V.
4.86%
20.6%
9.38%
12.0%
2.88%
7.53%
Mean
4.40
0.323
31.3
0.722
1.281
0.141
45°
Sn-1
0.217
0.020
3.5
0.027
0.047
0.005
C.V.
4.94%
6.09%
11.1%
3.71%
3.64%
3.66%
Mean
4.39
0.322
20.7
0.471
1.318
0.138
90°
Sn-1
0.178
0.021
4.4
0.069
0.022
0.007
C.V.
4.05%
6.52%
21.2%
14.6%
1.63%
4.80%
Manuscript preprint
Table 4: Measured material density of Nomex honeycomb cores.
Manufacturer
Nominal cell size (H) [mm]
ρcore [kg/m3]
Core height (T) [mm]
ρmat [g/cc]
tth [mm]
McGill
3.18
51.7
8.85
1.114
0.0553
Excel
3.18
76.6
12.70
1.142
0.0800
Plascore
3.18
51.6
17.80
1.115
0.0552
McGill HD322
4.76
31.4
25.40
1.148
0.0489
McGill HD342
4.76
60.0
25.40
1.160
0.0924
Table 5: Core cell wall strain gage raw elastic modulus results.
Specimen
Tension [GPa]
{S.G. 1, S.G. 2}
Compression [GPa]
{S.G. 1, S.G. 2}
Specimen
Tension [GPa]
{S.G. 1, S.G. 2}
Compression [GPa]
{S.G. 1, S.G. 2}
32 kg/m3 - 1
{5.17, 4.86}
{4.75, 4.88}
64 kg/m3 - 1
{5.52, 6.56}
{4.91, 4.57}
32 kg/m3 - 2
{5.64, 5.78}
{5.28, 6.37}
64 kg/m3 - 2
{6.56, 4.99}
{6.11, 5.37}
32 kg/m3 - 3
{4.86, 6.01}
{5.73, 5.35}
64 kg/m3 - 3
{5.16, 4.83}
{5.56, 4.70}
Mean
5.39
5.39
Mean
5.60
5.20
Sn-1
0.49
0.60
Sn-1
0.78
0.59
C.V.
9.14%
11.0%
C.V.
13.9%
11.3%
Table 6: Nomex/phenolic calibrated elastic properties and test-simulation comparison.
Identified material properties
E1 [GPa]
E2 [GPa]
G12 [GPa]
ν12
5.20
3.22
1.26
0.24
Gage results
E [GPa]
E45° [GPa]
E90° [GPa]
ν0°-90°
ν45°
ν90°-0°
Test
5.31
4.40
4.39
0.292
0.323
0.322
Model
5.31
4.40
4.39
0.309
0.361
0.259
Table 7: Hexcel 3.18 mm 80 kg/m3 honeycomb core test and simulation properties.
Case
Compression
modulus [MPa]
L-direction shear
modulus [MPa]
W-direction shear
modulus [MPa]
1- Manufacturer test data
255
70.0
37.0
2- FEM with identified properties, φ = 47.3°
(Table 6)
219
52.9
21.8
3- FEM with calibrated properties, φ = 47.3°
(E1 = 5.20 GPa, E2 = 3.73 GPa, G12 = 1.85 GPa, ν12 = 0.24)
255
77.4
31.9
4- FEM with φ = 60°, ρ = 80 kg/m3, t2 = t1
(E1 = 5.20 GPa, E2 = 3.61 GPa, G12 = 1.79 GPa, ν12 = 0.24)
255
72.3
35.9
5- FEM with φ = 60°, ρ = 80 kg/m3, t2 = 1.77·t1
(E1 = 5.20 GPa, E2 = 3.60 GPa, G12 = 1.76 GPa, ν12 = 0.24)
255
70.1
36.8
6- FEM with φ = 60°, ρ = 80 kg/m3, t2 = 1.74·t1
(E1 = 5.20 GPa, E2 = 3.60 GPa, G12 = 1.76 GPa, ν12 = 0.24)
255
70.0
37.0
... Other studies have investigated potentially non-homogeneous biobased cores, especially HC cores with bio-based polymer matrices and/or bio-based fibre reinforcement. Paper-based HCs, such as the Kraft paper and Nomex HCs, are commercially known bio-based HC alternatives as sandwich panel core with good mechanical properties [2,67,68]. Other types of HC cores with bio-based reinforcement have been the focus of recent investigations. ...
Article
Sandwich panels are one of the most applied designs in the aerospace, construction and automotive structures due to their excellent stiffness-to-weight ratio. Recent advances in the design of eco-friendly sandwich structures with bio-based and/or recycled/recyclable components indicate a clear move towards reducing the environmental impacts of the resulting designs while preserving mechanical performance. This movement is driven by strict environmental legislation and the need to save material costs and fuel consumption, requiring the use of greener components in the transport and building sector with higher loading performance and a smaller environmental footprint. This work presents the most recent efforts in the development of eco-friendly sandwich structures by classifying different approaches to the environmental optimisation of structures. A consistent classification of sandwich panels as either eco-friendly, bio-based, or quasi/fully-green structures is proposed according to their amount of bio-based components. The main components of bio-based/green sandwich panel skins, cores, and adhesives are listed, identifying the most promising designs under different mechanical properties. The environmental benefits and limitations of bio-based/green components are balanced by considering the eco-efficiency of the structures and outlining the challenges that are expected in the future.
... Skin failure (facial yielding, intra-cell dimpling, and face wrinkling) and core failure are two common modes of failure (core shear and local indentation) (Petras, 1999). However most of the analysis used three point bending test and tensile test set-up, moreover, majority of the researcher's added boundary conditions on the honeycomb sandwich, treated the sandwich as simply supported beam (Birman & Kardomateas, 2018;Gao et al., 2020;Roy, Park, et al., 2014). A lot mimic the 3-point bending testing, adding two supports and a pusher. ...
Article
Honeycomb sandwich is really one of the fundamentals to make a composite strong, stiff, very light, safe and have wonderful performance. Honeycomb materials are majorly used where high strength to weight ratio, stiffness to weight ratio is needed. Honeycomb sandwich consist of two face sheet or skin and a light core which can take many shapes, the common is hexagonal shape. The core handles shear load, while the skins resist compression and tension. This paper aims to guide the design of honeycomb sandwich structures done with finite element analysis software. The characteristic of honeycomb at microstructure and unit cell will be discussed Moreover, much demand on light weight honeycomb structures that can withstand heavy loads under different working condition are on high demand. Experimental approach can be time consuming and costly, this created room for massive research using FEA on loading response with various cores and thickness, in order to investigate the mechanical properties. This study will focus on the FEA of honeycomb sandwich done by many researches currently on commercial software's ANSYS and ABAQUS, this will be a guideline for researches to see what has been done and what is obtainable using FEA software.
... In the present work, we expand on this work by investigating the use of polyaramid sheets for the fabrication of 3D carbon origami structures. We use Nomex (chemical name: poly (m-phenylenediamine isophthalamide)), a polyaramid typically used as a flame retardant material due to its excellent thermal, chemical, and radiation resistance properties [22,23] as well as reinforcing materials in several high-performance structural applications [22,24,25]. The carbonization of Nomex material has been studied by several researchers for the production of activated carbon materials, which have been demonstrated for applications including gas separation, catalysts, catalyst supports, and electrodes for energy storage devices [26][27][28][29]. ...
Article
Full-text available
Carbon origami enables the fabrication of lightweight and mechanically stiff 3D complex architectures of carbonaceous materials, which have a high potential to impact a wide range of applications positively. The precursor materials and their inherent microstructure play a crucial role in determining the properties of carbon origami structures. Here, non-porous polyaramid Nomex sheets and macroporous fibril cellulose sheets are explored as the precursor sheets for studying the effect of precursor nature and microstructure on the material and structural properties of the carbon origami structures. The fabrication process involves pre-creasing precursor sheets using a laser engraving process, followed by manual-folding and carbonization. The cellulose precursor experiences a severe structural shrinkage due to its macroporous fibril morphology, compared to the mostly non-porous morphology of Nomex-derived carbon. The morphological differences further yield a higher specific surface area for cellulose-derived carbon. However, Nomex results in more crystalline carbon than cellulose, featuring a turbostratic microstructure like glassy carbon. The combined effect of morphology and glass-like features leads to a high mechanical stiffness of 1.9 ± 0.2 MPa and specific modulus of 2.4 × 104 m2·s−2 for the Nomex-derived carbon Miura-ori structure, which are significantly higher than cellulose-derived carbon Miura-ori (elastic modulus = 504.7 ± 88.2 kPa; specific modulus = 1.2 × 104 m2·s−2) and other carbonaceous origami structures reported in the literature. The results presented here are promising to expand the material library for carbon origami, which will help in the choice of suitable precursor and carbon materials for specific applications.
... The skins are manufactured with prepreg plies with high modulus carbon [19,20]. The orientation of the material coordinate system corresponds to that usually used for this type of anisotropic material [21]. ...
Article
Full-text available
A new testing set-up to measure the compressive strength of continuous fibre composite materials (CFRP) under both static and fatigue loadings is described. A specific symmetric sandwich beam with CFRP face sheets is used with a four-points bending set-up. Its mechanical design is carried out by fine Finite Element Analyses (FEA) to ensure that the failure mode is axial compression by fibre micro-buckling. Precise comparisons between numerical simulations and experimental quasi-static experimental results are found to be sound, validating the FEA model. The latter is used to design the sandwich beam in terms of strength and stability criteria. Finally, a high cycle fatigue campaign (107 cycles) with high modulus CFRP skins demonstrates the performance of the set-up. Repeatable results and acceptable failure modes are obtained.
... CFRP is one of the most ideal materials in the fields of lightweight design and impact protection due to its excellent physical properties [4,5]. Honeycomb core with different geometries, materials, arrangements, and loading conditions possess a series of excellent properties which has been became an important research object for weight-reducing and energy-absorbing [6][7][8]. Sandwich structures with composite skin and porous cores have become a significant structure for lightweight and rigidity design [9][10][11][12][13][14]. On the other hand, many studies manifest that the engine hoods made of composite materials exhibit better properties than the traditional hoods in the aspect of weight reduction and impact protection [15][16][17]. ...
Article
Honeycomb sandwich structures possess many excellent characteristics such as light weight, high stiffness, high modulus and good vibrate isolation, which have became important structures for energy saving and emission reduction. However, most of the researches are concentrated on flat honeycomb sandwich structures and few of that considers the curved structures. In this paper, double-curved honeycomb sandwich core is investigated by mechanical analysis, and equivalent mechanical model is established. Detailed models of the honeycomb sandwich structure are established. Simultaneously, the corresponding equivalent models are built on the sandwich theory and curved sandwich theory, respectively. Afterward, a double-curved sandwich hood consisting of two carbon fiber reinforced plastic (CFRP) skin and Nomex honeycomb core was designed and optimized based on equivalent model and a traditional aluminum alloy engine hood. The whole optimization progress composed of free-size optimization, size optimization and stacking sequence optimization. The optimized structure with a weight reduction by 20.1% while maintaining the original rigidity. In addition, compared with the referenced aluminium alloy engine hood structure, the weight of the optimal structure was reduced 61.8% on the condition of better stiffness performance.
Article
Sandwich structures are frequently used in structural areas where lightness and strength are essential. These structures are indispensable for sailing boats, and ground and air vehicles. The base purpose of this study is to investigate the effect of wave parameters on the sandwich structure. The data obtained from the bending tests of the model created using Ls-Dyna was compared with the experimental data of the literature. There is a 3.05% difference between the peak force in experimental and Ls-Dyna. The force-deformation plots are coherent, and the progressive images of the sandwich structure during bending are similar. In addition, using theoretical approaches, the highest force and the amount of collapse during bending were determined. There is a difference of 3.1% between the theoretical approach and Ls-Dyna values. Thus, the Ls-Dyna model was validated. The flat cell walls of the honeycomb were modeled as a sine wave. Four wave numbers and wave amplitudes were used. In this way, 16 different analysis files were created. The results show that the new sandwich structure’s specific peak force and specific energy absorption (SEA) increased by 7–110% compared to the ordinary flat walled sandwich structure. This research will assist in the design of new sandwich structures.
Article
A progressive damage model for aramid honeycomb cutting was proposed to reveal its cutting damage mechanism. It established the relationship between the mesoscale failure modes and the macroscale cutting damage types of the aramid honeycomb. The proposed model addressed the material assignment problem of impregnated honeycomb by developing a material calculation method that simulates the real manufacturing process of the aramid honeycomb. Cutting experiment of aramid honeycomb specimen was conducted concerning on the cutting forces response and cutting damages, which validated that the proposed method was effective for investigating the cutting process and mechanism for the aramid honeycomb. Predicted cutting mechanism results show that: (a) cutting process of the aramid honeycomb can be divided into three stages with four characteristic states—initial state, cut-in state, cut-out state and final state; (b) cell wall bending in the cutting direction relieves the cutting force, and strong plasticity of the aramid fiber makes it hard to break, which lead to uncut fiber and burr damages; (c) using sharp tip cutting tool to reduce cutting force and bonding both top and bottom of the honeycomb to make it stiffer are beneficial to obtain good cutting quality with less damages.
Article
A comprehensive investigation is presented on the Fused Filament Fabrication (FFF) technology’s possibilities to create cellular solids with a broad spectrum of specific stiffness and strength, modifying cell geometry and size, while addressing manufacturing matters such as inherent defects and built time. Thirteen typologies of two-dimensional cellular patterns with different relative densities are examined. Results have allowed conclusions to be drawn regarding the influence of cell type and infill density on mechanical performance. Intra-layer and inter-layer inherent defects identified after manufacturing highlight the importance of optimizing filament trajectories. A reliable comparison of the elastic properties of the cellular patterns as a function of their density is presented. An experimentally validated numerical model is provided for predicting the compression stiffness of the different cell patterns with an average deviation below 5%. The model can reproduce the behavior in the elastic range based on tensile specimen properties and a Normal Stiffness Factor to account for the phenomenon of elastic asymmetry of the FFF print samples. The wide range of results achieved is experimental confirmation of the potential of FFF cellular solids. Lastly, this investigation provides analytical, numerical, and empirical validated evidence to further design-for-additive manufacturing strategies with cellular solids for designing advanced lightweight structures.
Article
To address the problem of the poor stability of ultrasonic machining of wave-absorbing honeycomb material, this study takes ultrasonic cutting of wave-absorbing honeycomb material with a disc cutter as the research object and establishes a multi-degree-of-freedom mathematical model of the cutting system based on the relative positions of the tool, the wave-absorbing honeycomb material, and the motion characteristics of the tool. On this basis, modal analysis of the disc cutter and the honeycomb cell wall plate is carried out to draw the Lobe diagram of ultrasonic cutting stability, the process experimental parameters are determined to accord to the solved stability Lobe diagram, and machining stability verification experiments are carried out. The experimental results show that the machining parameters in the stable zone of the Lobe diagram result in a neat and clean surface, less fibre pullout, a complete outer substrate, and less tool wear than those in the critical and unstable zones, thus verifying the correctness of the theoretical model and the stability Lobe diagram.
Conference Paper
Full-text available
Driven by stringent weight saving requirements composite sandwich construction has evolved as one of the basic structural design concepts for load-carrying components of advanced aeroplanes and helicopters. Particularly, sandwich using laminated carbon fibre reinforced plastics (CFRP) as face sheets and NOMEX honeycombs as core material is increasingly used due to features such as high strength-to-weight and stiffness-to-weight ratios as well as an excellent fatigue behaviour. While offering unique advantages, sandwich is also prone to a range of defects and damages. Due to the thin brittle skins and the weak core material CFRP sandwich structures are particularly susceptible to impact loading which may accidentally occur during assembly or operation of aircraft. Since these damages may have detrimental effects on the load carrying capability, they have to be considered in the damage tolerant design of aircraft structures. For that purpose it is necessary to determine the extent of damage in sandwich structures, resulting from impact events such as tool drop or thrown up debris. Up to now this task is mainly done experimentally by using drop weights to simulate the impact loading and NDT methods to determine the damage size. Since these experimental procedures are rather costly and time consuming, there is a clear need to supplement them by reliable numerical simulation tools. Usually, explicit finite element methods are employed for this task. As far as the global behaviour of sandwich components is investigated by finite element methods it is sufficient to model the structure by using shell elements for the skins and solid elements for the core. In the case of honeycombs such models permit only a macro-mechanical description of the core behaviour. Thus, it is not possible to account for local failure modes of the hexagonal cell structure. Nevertheless, these local effects are important, if impact loading is investigated. For this kind of problem a detailed modelling of the honeycomb cell structure is required [1, 2]. Impact loading of honeycomb sandwich results in very complex damage modes in the core as can be seen in Figure 1a. For example, in the centre of the impact area the honeycomb material is crushed. This damage is mainly the result of local buckling of the cell walls and compressive failure of the resin-impregnated NOMEX paper. Closer to the edge of the impact area the core is subjected to high shear forces which results in shear cracks in the cell walls. This clearly shows that material properties such as stiffness and strength of the honeycomb material are crucial parameters for the formation of damage in the sandwich. Therefore, the knowledge of these properties are essential for a reliable numerical simulation of impact loaded honeycomb sandwich structures as shown in Figure 1b. Usually, honeycomb manufacturers provide material data only for honeycomb blocks instead of the impregnated papers used as core material. So, one approach is to determine the paper properties from the global core properties. This can be done by numerical simulation of the tests used to measure the global properties [1]. This approach provides only averaged data and it is difficult to identify the non-linear behaviour of the material. Therefore, a research project was initiated which aimed at the determination of material properties of resin-impregnated NOMEX paper. Two main problems had to be solved. The first one was to get sheets of impregnated paper with a sufficient size for testing. Since the hexagons of the honeycombs usually applied in aircraft structures are very small, it is not possible to cut test specimens directly from the cell walls. Therefore, larger sheets of NOMEX papers had to be produced by a honeycomb manufacturer. The applied process was similar to that used for the production of standard honeycomb cores. The second problem was to find appropriate experimental methods. In a first move, standard test methods were evaluated [e.g. 3, 4, 5] which are normally used to determine material properties of paper and board. The result of this evaluation was that only the standard test for tensile properties [3] is applicable. Therefore, two new tests have been developed. The first one is used to measure the mechanical behaviour of impregnated NOMEX under compression (see Figure 2, left). It is based on the ring crush method of ISO 12192 which was modified to prevent premature buckling of the test specimen and to provide well defined clamping conditions. The second test fixture was designed for the determination of the behaviour of NOMEX paper under shear loading (see Figure 2, right). For this test fixture the concept of rail shear tests was chosen and adapted to the special requirements of the thin walled material.
Article
Full-text available
Nomex™ honeycomb core composite sandwich panels are widely used in aircraft structures. Detailed meso-scale finite element modeling of the honeycomb geometry can be used to analyze sandwich inserts, vibration response, and complex combined loading cases. The accuracy of a meso-scale honeycomb modeling technique for static load cases was evaluated. A rectangular honeycomb core was modeled with perfect hexagon honeycomb cells. Compression and shear tests simulations with linear and non-linear solutions were performed for four core densities. The simulated moduli and buckling strengths were recorded. These results were compared to property data published by honeycomb manufacturers. The simulated maximum honeycomb wall stresses at the manufacturer predicted core strengths were also recorded. The honeycomb walls’ first compression deformation mode shape was observed. Sinusoidal small imperfections were then introduced in the honeycomb geometry based on that deformation mode shape. These imperfections provided a better match to manufacturer compressive modulus data while having a limited impact on the shear moduli. The simulated properties did not exactly match manufacturers’ shear and compression data together for all the core densities. Modeling the honeycomb cells with rounded corners and with increased thickness at the cell junctions are potential strategies to improve the accuracy.
Article
The present study investigated fracture and various mechanical properties of phenolic resin. It appears that while in the linear elastic region, measured modulus is independent of loading mode, the ultimate strengths differ considerably in the three types of loading used: tension, three point bending, and compression. The concepts of linear elastic fracture mechanics (LEFM) are applied to characterize the fracture behaviour of phenolic resin in terms of fracture toughness KIC and GIC respectively under slow and impact loading using pre-notched test specimens in three point bending. It is found that KIC is an adequate criterion for fracture and may be related to G, by using the Irwin-Kies formula.
Article
Within the framework of research work carried out at the Institute of Papermaking and Printing, the Technical University of Łódź, Poisson's ratios were determined in a paper plane. This feature presents a simple and easy method how to determine this material constant. The method proposed for the determination of Poisson's ratio utilises the results of typical procedures used when defining physical properties, such as Young's modulus, paper density and the TSI (tensile stiffness index). For practical verification of the method proposed, the author tested papers for the production of corrugated board and compared the values of ratios obtained for the propagation velocity of sonic waves in the machine and cross directions.
Article
The compression strength and the shear elastic modulus of an aramid honeycomb core have been studied theoretically and experimentally. Equations have been proposed for calculation of these characteristics based on the longitudinal and shear elastic moduli of the paper material composing the core and the detailed geometry of the honeycomb shape. The effects of single and double wall thickness of the materials are taken into account in the equation. The predicted mechanical properties showed good agreement with the test results although a difficulty was found in the determination of the shear elastic modulus of the paper material. It also became clear that the aramid papers used showed considerable anisotropy which seemed to strongly influence the shear property of the core.
Article
An experimental study has been conducted to investigate the effects of elevated temperature and humidity on the strength and failure mode of carbon/BMI-Nomex composite sandwich joints. Pull-out and shear tests were performed under three different humidity conditions (dry, pre-saturation, and full saturation) and five different temperature conditions (−55, 24, 82, 177, and 204 °C). With respect to pull-out loading, while the ultimate failure modes were diverse in accordance with the environmental conditions, the initial damage to the joints was always due to core shear buckling. In the joints under shear loading, the bearing failure of the upper face by the insert flange and insert separation from the potting mass commonly occurred as the main failure modes under every environmental condition. The pull-out and shear strengths of the sandwich joints at the coldest temperature (−55 °C) in the dry condition always showed the highest values among all conditions.
Article
In nearly all sandwich constructions certain types of joints have to be used for assembly, but little is known about their failure behaviour. This paper deals with the investigation of the mechanical behaviour of three different corner joints as a right-angled connection of two sandwich panels and of two different potted inserts as a localised load introduction in Nomex® honeycomb sandwich structures with glass fibre-reinforced composite skins. For this purpose, experimental test series were conducted including shear tests and bending tests of the corner joints and pull-out as well as shear-out tests of the threaded inserts. The failure mechanisms and sequences are described for each load case and the influence of the different designs and of the loading rate is discussed. Based on these characteristics, finite element simulation models were developed in LS-DYNA, which are able to represent the respective failure behaviours.
Article
Virtual testing using dynamic finite element simulations is an efficient way to investigate the mechanical behaviour of small- and large-scale structures reducing time- and cost-expensive prototype tests. Furthermore, numerical models allow for efficient parameter studies or optimisations. One example, which is the focus of this paper, is the configurational design of cellular sandwich core structures. From classical honeycomb cores to innovative folded core structures, a relatively large design space is provided allowing for tailoring of the cellular core geometry with respect to the desired properties. The method of determining the effective mechanical properties of such cellular sandwich core structures of different geometries using dynamic compression, tensile and shear test simulations is discussed covering a number of important modelling aspects: the cell wall material modelling, the influence of mesh size and number of unit cells, the inclusion of imperfections, etc. A comparison of numerical and experimental results is given for Nomex® honeycomb cores and Kevlar® or carbon fibre-reinforced plastic (CFRP) foldcore structures. A good correlation with respect to cell wall deformation mechanisms and stress–strain data was obtained. Therefore, these models not only allow for a complete mechanical characterisation of cellular core structures but also for a detailed investigation of cell wall deformation patterns and failure modes to get a better understanding of the structural behaviour, which can be difficult using solely experimental observations. To show that this efficient virtual testing method is suitable for the development of cellular core geometries for specific requirements, an optimisation study of a CFRP foldcore geometry with respect to its compressive behaviour was performed.