Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 88 (2014) 149 – 156

International Symposium on Dynamic Response and Failure of Composite Materials, DRaF2014

Bearing mode absorber – On the energy absorption capability of

pulling a bolt through a composite or sandwich plate
Sebastian Heimbs*, Tim Bergmann
Airbus Group Innovations, 81663 Munich, Germany

Abstract

An experimental study was conducted to investigate the energy absorption capability of continuously pulling a metallic bolt
through a composite or sandwich plate. The influence of various parameters like the fibre and matrix material as well as the fibre
architecture and angle on the specific energy absorption under quasi-static and high-rate dynamic loading was assessed. The
achievable quasi-static weight-specific energy absorption (SEA) values up to 164 kJ/kg are very promising compared to classical
absorbers. Considerable strain rate effects led to significantly lower values under high-rate loading, though. The thin-skinned
sandwich specimens used in this study were not competitive with lower SEA values.
©
© 2014
2014 The Authors. Published
The Authors. PublishedbybyElsevier
ElsevierLtd.
Ltd.This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the Organizing Committee of DRaF2014.
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Organizing Committee of DRaF2014
Keywords: Composite material; Sandwich; Bearing; Energy absorber; Loading rate effect

1. Introduction

Energy absorbers are used in numerous occupational safety or transport applications like in passenger cars,
railway or aircraft structures for collision or crash energy absorption. Their purpose is to reduce the loads or
decelerations acting on a vulnerable component (e.g. the passenger) typically by controlled deformation or
destruction of an energy absorbing structure. In this context, the crushing and fragmentation of fibre-reinforced
polymer matrix materials (FRP) is well known for its high weight-specific energy absorption capability (SEA) [1].
Consequently, composite crushing absorbers are applied today in many high-performance vehicle structures like in
sports cars, high-speed trains or helicopters.

* Corresponding author. Tel.: +49-89-607-25884; fax: +49-89-607-23067.

E-mail address: sebastian.heimbs@eads.net

1877-7058 © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Organizing Committee of DRaF2014
doi:10.1016/j.proeng.2014.11.138
150 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156

Nomenclature

AF Aramid fibre
CF Carbon fibre
EP Epoxy resin
FRP Fibre-reinforced polymer composite
GF Glass fibre
LCP Liquid crystal polymer fibre
PA Polyamid
PEEK Polyetheretherketone
PF Phenolic resin
UHMWPE Ultra-high-molecular-weight polyethylene fibre
SEA Weight-specific energy absorption
SF Steel fibre

Although many such energy absorber solutions have been developed for compressive loads, several load cases
require tensile absorbers that absorb energy under tension. Classical examples are fall arrest absorbers for persons or
items [2,3,4]. One promising option to use composite materials for tensile absorbers is to pull a bolt in a bearing
mode through a composite plate, which enables a relatively constant load level that is beneficial for efficient energy
absorption and high SEA values. This concept was initially investigated as an option for tensile absorbers for
overhead stowage bins in aircraft cabins [5,6,7]. Its suitability for aircraft seat absorbers was addressed in [8,9].
Other publications can be found with similar tests that aim at characterising the residual strength after initial bearing
failure without the explicit aim to use this mechanism as an energy absorber [11,11,12].
Because only very few data exist on the energy absorption capability of the bearing failure mode of a bolt being
continuously pulled through a composite or a sandwich plate, an extensive test campaign was conducted in the
framework of this study. The aim was to assess in a comprehensive manner the influence of fibre and matrix
material as well as the fibre architecture and angle on the energy absorption capability under quasi-static and high-
rate dynamic loading.

2. Testing procedure

A special clamping device was developed that allows for the clamping of composite and sandwich plates with a
size of 150 mm x 75 mm via six fixed bolts (Fig. 1) [13]. A steel bolt with a diameter of 8 mm was chosen to be
pulled through the plate with the results of this configuration being directly comparable to the studies in [5,6,7].

Fig. 1. Image sequence of bolt pull-through test (left) and post-test specimen view (right).
 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156 151

All quasi-static tests were performed at room temperature on a Zwick universal testing machine with a loading
rate of 200 mm/min. Additional high-rate dynamic tests with a loading rate of 3 m/s were performed on a servo-
hydraulic Zwick HTM testing machine in order to evaluate potential strain rate effects and to expose the absorber to
realistic loading rates of a highly dynamic energy absorption event.
The key indicator of the energy absorber performance is the SEA value, which is calculated by dividing the
absorbed energy EA (area under the force-displacement curve) by the mass of the destroyed material (Eq. 1). The
destroyed material volume is defined as 1.2 (brittle fibres) or 1.5 (ductile fibres) times more than the basic slot
volume; which is the plate thickness t multiplied by the bolt diameter d and the maximum displacement smax (Fig. 2).
This conservative definition is based on micrographs of the lateral extent of damage in the composite plates, which is
typically 20% or 50% higher than the bolt diameter for brittle and ductile fibres, respectively. It is worth mentioning
that this SEA value is defined on material level for comparison of material properties. An alternative definition is the
SEA based on final energy absorber level, which takes the mass of all material into account that is necessary to
achieve the absorber performance, incl. the surrounding material, support structures and load introduction devices.

s
EA 1
SEA = = F (s )ds (1)
mabsorbed ρ t d smax ∫0

Force F EAideal

Metallic bolt
smax FRP material structure

Material for energy absorption EA

d Displacement s

Fig. 2. Schematic of a bearing mode energy absorber and typical force-displacement response.

3. Test results of composite plates

The first test campaign was conducted with monolithic composite plates with various fibre and matrix systems.
The plate thickness was between 1.3 and 2.7 mm and the 8 mm hole was chamfered from both sides to act as a
trigger and to eliminate any initial peak load (Fig. 3). Table 1 gives an overview of the fibre and matrix
combinations, referred to as specimen types A to M.

150 mm

100 mm bolt pull-through length Ø8 mm

75 mm

Ø8 mm bolt hole (sym. chamfered)

6 x M6 clearance holes

Fig. 3. Dimensions of the composite specimens.

152 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156

Table 1. Overview of composite fibre/matrix combinations tested.

Type Fibre Matrix Density
[g/cm³]
A CF HTA carbon fibre fabric (plain weave, 200 g/m²) EP Hexcel RTM6 epoxy resin 1.45
B CF HTA carbon fibre fabric (plain weave, 200 g/m²) EP Cytec EP2400 toughened epoxy resin 1.50
C CF HTA carbon fibre fabric (plain weave, 200 g/m²) PEEK Thermoplastic polyetheretherketone 1.55
D CF HTS carbon fibre 3D weave (2720 g/m²) EP Hexcel RTM6 epoxy resin 1.50
E CF HTS carbon fibre UD [60°/90°/-60°/60°,90°/-60°]s EP Cytec 977-2 toughened epoxy prepreg 1.58
F CF T800S carbon fibre UD [45°/90°/-45°/45°/45°/0°/90°]s EP Hexcel M21 toughened epoxy prepreg 1.53
G GF E-Glass fibre fabric (plain weave, 390 g/m²) EP Hexcel RTM6 epoxy resin 1.77
H GF E-Glass fibre fabric (satin weave 1/4) PA Thermoplastic polyamid 1.67
I GF E-Glass fibre fabric (satin weave 1/8, 296 g/m²) PF Phenolic resin prepreg 1.77
J AF Twaron® 2200 fabric (plain weave, 170 g/m²) EP Hexcel RTM6 epoxy resin 1.30
K LCP Vectran® fabric (twill weave 2/2, 200 g/m²) EP Hexcel RTM6 epoxy resin 1.28
L UHMWPE Dyneema® fabric (plain weave, 130 g/m²) EP Epikote L20 epoxy resin 1.06
M SF Stainless steel fibre fabric (plain weave, 1195 g/m²) EP Hexcel RTM6 epoxy resin 4.93

3.1. Influence of fibre material

The comparison of different fibre materials in a similar plain weave fabric configuration in 0°/90° orientation
with epoxy resin under quasi-static loading led to highest SEA values for glass (G) and HTA carbon (A) fibres,
followed by Dyneema® (L), aramid (J) and Vectran® (K) fibres. Stainless steel fibres (Bekaert FA900, ∅ 12 µm)
(M) with their high density and low strength were not competitive with very low SEA values.

3.2. Influence of resin material

Switching from relatively brittle RTM6 epoxy resin to toughened epoxy resin EP2400 led to a strong increase of
about 18 % for 0°/90° and ±45° orientation. Using a thermoplastic matrix such as PEEK in combination with carbon
fibres resulted in comparable SEA values as for the toughened epoxy resin. For glass fibre materials the substitution
of RTM6 epoxy resin with PA6 had a negligible effect on the SEA values, whereas phenolic resin showed by far the
weakest performance resulting in lowest SEA values.

3.3. Influence of unidirectional vs. 2D fabric vs. 3D fabric

Specimens with a lay-up of unidirectional carbon fibre plies and comparable untoughend and toughened epoxy
resin materials showed 15-30 % higher SEA values than woven fabric specimens under 0°/90° orientation, even
though the mass mabsorbed was increased by 10% because of the extensive fibre pull-out at the top and bottom surface
of the unidirectional specimens. The unidirectional plies consisted of high strength HTS fibres, though. The use of
3D-woven fabrics led to no improvement, because no stable bolt pull-through behaviour could be achieved.

3.4. Influence of fibre orientation angle

Changing the fabric orientation angle from 0°/90° to ±45° relative to the bolt direction led to an increase in SEA
values of about 10-20 % for carbon fibres (A, B and C). This can be related to the highest tensile forces acting in
±45° relative to the bolt direction and therefore correlates to the fibre direction [6]. However, the stress displacement
curves for the ±45° orientation are typically more oscillating and less uniform and reproducible. A similar
comparison of 0°/90° and ±45° orientation for configurations with ductile fibres (J, K and L) was not possible,
because none of the configurations failed in a controlled and reproducible manner.
 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156 153

3.5. Influence of loading rate

The loading rate had a significant influence on the energy absorption capability of the tested material
configurations (Table 2, Fig. 4). The change in loading rate from quasi-static (200 mm/min) to high-rate dynamic
loading (3 m/s) resulted in a significant decrease in crushing stress and hence a reduction in SEA values by 20-40 %.
This is in agreement with literature data of dynamic crushing of composite structures which also showed 20-40 %
reduction in crushing stress levels at comparable or even higher test velocities [7,12,14,15,16].

3.6. Overview of test results

An overview of the experimental results in terms of crushing stress (mean force devided by the product of plate
thickness and bolt diameter) and SEA values is given in Table 2. All in all, SEA values of 24-164 kJ/kg could be
achieved. The highest quasi-static value was obtained with specimens B (CF fabric with toughened epoxy resin
under ±45°), for the dynamic load case with specimens F (unidirectional CF with toughened epoxy resin).

Table. 2. Overview of failure stress and SEA values on material level for the tested configurations.
Type Conf. Mean crushing stress [MPa] SEA [kJ/kg]
Quasi-static Dynamic Quasi-static Dynamic
A 0°/90° 194 144 112 83
±45° 239 155 137 95
B 0°/90° 239 145 133 81
±45° 295 173 164 96
C 0°/90° 241 146 129 88
±45° 282 173 152 93
D 0°/90° 441 - 98 -
E 311 - 146 -
F 316 237 150 125
G 0°/90° 291 202 137 97
±45° 279 222 131 105
H 0°/90° 265 - 132 -
±45° 242 - 121 -
I 0°/90° 115 140 54 66
±45° 152 144 71 68
J 0°/90° 206 173 106 89
±45° 307 - 63 -
K 0°/90° 180 105 94 55
±45° 292 - 61 -
L 0°/90° 171 105 107 66
±45° 366 - 92 -
M 0°/90° 214 - 24 -
±45° 210 - 24 -
N L-dir. 12 13 50 53
W-dir. 14 14 58 59
O L-dir. 29 29 68 66
W-dir. 31 28 73 65
P L-dir. 34 32 79 75
W-dir. 33 32 75 75
154 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156

Fig. 4 gives an impression of the post-test view of different specimens as well as micrographs of the crushing
zone in front of the metallic bolt. Lamina bending and buckling, fibre and matrix fracture as well as delaminations
can be observed as the major failure modes. The lateral extent of damage for the exemplary specimen of type C is
given in Fig. 5. The representative curves of stress vs. bolt displacement for quasi-static and high-rate dynamic
loading in Fig. 4 highlight the significant reduction of crushing stress.

CF/EP (B): CF/PEEK (C): GF/EP (G): GF/PF (I):

quasi-static

dynamic

AF/EP (J): Vectran®/EP (K): Dyneema®/EP (L): SF/EP (M):

Fig. 4. Test results and micrographs of composite specimens (in 0°/90° direction).
 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156 155

Fig. 5. Lateral extent of damage for specimen C is approx. 20% larger than the bolt diameter.

4. Test results of sandwich plates

The test campaign was extended to sandwich plates with phenolic-impregnated aramid paper honeycomb cores
and glass fibre-reinforced skins of different thicknesses (referred to as N, O and P). The total sandwich thicknesses
ranged from 9.5 to 13 mm, the skin thicknesses from 0.4 to 1.6 mm. The lateral core material had partially to be
removed and replaced by wooden ledges to enable the clamping of the specimen edges. The achievable SEA values
of 50-79 kJ/kg were much lower than the ones of the monolithic plates. The honeycomb core orientation in L- or W-
direction as well as the loading rate had no significant influence on the test results, which can be seen in Fig. 6.

0.4 mm GF/EP + 9 mm honeycomb (N): 1 mm GF/EP + 9 mm honeycomb (O): 1.6 mm GF/EP + 11 mm honeycomb (P):

Fig. 6. Test results and micrographs of sandwich specimens (in L-direction).

156 Sebastian Heimbs and Tim Bergmann / Procedia Engineering 88 (2014) 149 – 156

5. Conclusion

An extensive experimental study was conducted to investigate the energy absorption capability of bearing mode
failure by continuously pulling a bolt through a composite plate. The achievable quasi-static SEA values on material
level of up to164 kJ/kg are very promising compared to classical absorbers [1]. Considerable strain rate effects led to
significantly lower values under high-rate loading, though, with a maximum value of 125 kJ/kg in these tests.
The investigation of different fibre and matrix materials as well as reinforcement configurations led to the
conclusion that the fibre strength appears to play a major role since high strength reinforcements showed the highest
SEA values. Highly ductile fibres had no major advantage. Toughened epoxy resins showed superior performance
compared to brittle resins. The unidirectional laminates led to higher SEA values compared to the woven fibre
fabric, which can be caused by the different fibre strength values using HTS and T800S fibres in comparison to HTA
fibre fabrics. Care should be taken when comparing these results. The study in [7] compared identical fibres in
unidirectional and woven fabric architecture and identified woven fabrics to be preferable. The sandwich specimens
were not competitive with relatively low SEA values.
One further factor, which is believed to have an influence on the test results, is the plate thickness with its
geometrical effect on the failure mechanisms. This was not investigated here and is the subject of further research.

Acknowledgements

This study was performed within the LuFo IV-4 project INCCA, funded by the German Federal Ministry for
Economic Affairs and Energy (BMWi). The financial support is gratefully acknowledged. Additionally, the authors
wish to thank Kurt Pfeffer from Airbus Group Innovations for performing the quasi-static tests and Sebastian
Schmeer and Stefan Gabriel from IVW Kaiserslautern for the dynamic testing activities.

References

[1] G.C. Jacob, J.F. Fellers, S. Simunovic, J.M. Starbuck, Energy absorption in polymer composites for automotive crashworthiness, J. Compos.
Mater. 36 (2002) 813-850.
[2] A.B. Spierings, R. Stämpfli, Methodology for the development of an energy absorber: application to worker security ropes, Int. J. Impact Eng.
32 (2006) 1370–1383.
[3] Y.M. Goh, P.E.D. Love, Adequacy of personal fall arrest energy absorbers in relation to heavy workers, Safety Sci. 48 (2010) 747–754.
[4] K. Baszczynski, Influence of weather conditions on the performance of energy absorbers and guided type fall arresters on a flexible
anchorage line during fall arresting, Safety Sci. 42 (2004) 519–536.
[5] M. Pein, D. Krause, S. Heimbs, P. Middendorf, Innovative energy-absorbing concept for aircraft cabin interior, International Workshop on
Aircraft System Technologies (AST 2007), Hamburg, 2007.
[6] M. Pein, D. Krause, S. Heimbs, P. Middendorf, Hybrid composite materials for a highly integrated energy-absorbing concept for aircraft
cabin interior, 5th Triennial International Aircraft Fire and Cabin Safety Research Conference, Atlantic City, NJ, 2007.
[7] M. Pein, Entwicklungsmethode für Hochleistungswerkstoffe am Beispiel von Energieabsorbern für Flugzeugkabinen, Phd thesis, TU
Hamburg-Harburg, 2009.
[8] C. Olschinka, A. Schumacher, Dynamic Simulation of Flight Passenger Seats, 5th German LS-Dyna Forum, Ulm, 2006.
[9] W. Extra, Mechanische Absorptionsvorrichtung, Patent doc. DE 199 26 085 A1, 2000.
[10] M. Postec, E. Deletombe, D. Delsart, D. Coutellier, Study of the influence of the number of inter-ply interfaces on the bearing rupture of
rivited composite assemblies, Compos. Struct. 84 (2008) 99–113.
[11] Y. Pekbey, The bearing strength and failure behavior of bolted E-glass/epoxy composite joints, Mech. Compos. Mater. 44 (2008) 397–414.
[12] G.S. Ger, K. Kawata, M. Itabashi, Dynamic tensile strength of composite laminate joints fastened mechanically, Theor. Appl. Fract. Mech.
24 (1996) 147–155.
[13] T. Bergmann, S. Heimbs, G. Tremmel, M. Maier, Investigation of a composite tensile energy absorption element under static and dynamic
loading, 16th European Conference on Composiute Materials (ECCM16), Seville, 2014.
[14] D. Dormegnie, D. Coutellier, D. Delsart, E. Deletombe, Studies of scale effects for crash on laminated structures, Appl. Compos. Mater. 10
(2003) 49–61.
[15] A.G. Mamalis, Y.B. Yuan, G.L. Viegelahn, Collapse of thin-wall composite sections subjected to high speed axial loading, Int. J. Vehicle
Des. 13 (1992) 564–579.
[16] P.H. Thornton, J.J. Harwood, P. Beardmore, Fiber-reinforced plastic composites for energy absorption purposes, Compos. Sci. Technol. 24
(1995) 275–298.