ID-1559
AN EXPERIMENTAL INVESTIGATION OF THE
BEARING FAILURE MECHANISMS IN BOLTED
COMPOSITE JOINTS
Yi Xiao and Takashi Ishikawa
Structures Division, National Aerospace Laboratory
6-13-1, Osawa, Mitaka, Tokyo, 181-0015, Japan
ABSTRACT: This paper presents an experimental investigation of the behavior in
mechanically fastened composite structures. CF/PIXA and CF/Epoxy material systems were
selected for the strength comparison, and a simple and highly functional measurement system
using a non-contact Electro-Optical Extensometer was adopted to evaluate the bearing
strength. A static tensile test for a double lap bolted joint was performed. During tests, the
fracture process under the load-deformation response was monitored using the acoustic
emission measurement. Some of the specimens were unloaded at different loading levels, in
order to assess the damage progress and failure mechanisms using soft X-rays and SEM
observation of the internal damage. Based on these experimental observations, it can be
concluded that the kink band, delamination, fiber-matrix shearing, and fiber-matrix splitting
failure formed the major failure modes of the shear cracks. The washer area and bolt clamping
conditions should influence the damage progress.
Keywords: CF/PIXA Composite, Bearing Strength, Failure Mechanism, Mechanically
Fastened Joint, Out-of-plane Deformation, Microscopic Damage Behavior
INTRODUCTION
Degradation and failure of aircraft structures are frequently initiated at the joints, so that
safety, durability and reparability are strongly influenced by the adequacy of joint design. The
next generation supersonic transport (SST) aircraft make extensive use of advanced
composite materials in both primary and secondary structure, and the mechanically fastened
joints have been used.
Mechanically fastened joints have generally been used for highly-loaded composite
components, although the low bearing stiffness of composites can lead to problems of bolt
bending and hole elongation under compressive loading. Significant research [1-4] has been
carried out for fastener joints, depending on the specimen geometry and fiber orientation, the
joint would be expected to eventually fail in a variety of modes, namely net-tension,
shearing-out, and bearing. From the viewpoint of structural design, it is preferred that bearing
failure in which a stable failure mode develops should occur. However, bearing failure is a
local compressive failure mode due to contact and frictional forces acting on the surface of
hole [5]. This fracture process is very complicated and is influenced by many parameters such
as washer dimension and lateral clamping [6-9].
Eriksson [6] has shown that several important parameters influence the bearing strength
including the lateral constraint condition and ply orientation. Wang et al. [7] examined the
bearing failure mechanism as a function of the clamping pressure using the bearing response
and bearing strength of bolted joints. In order to evaluate the bearing damage, a pin-loaded
bearing (without lateral clamping) and a bolted bearing (with lateral clamping) test was
conducted. Wu and Sun [8] investigated that the behavior of pin-contact failure in composite
laminates and found fiber micro-buckling in the 0o-plies of the laminate to be an important
role in the initiation of bearing damage. Camanho and Matthews [9] carried out a detailed
experimental investigation into three basic failure modes of the joint, with the accumulation
of delamination damage being the main mechanism of bearing failure. From the above
literature, the damage phenomenon of bearing failure in a bolted composite joint would
appear to be complicated. One of these complicated results is due to out-of-plane compressive
deformation in the vicinity of the hole. Therefore, the fiber microbuckling, matrix cracks and
delamination, etc. are easy to occur due to compressive load in laminate structures. An
understanding of the occurrence and progress of the microscopic damage behavior is thus
necessary to the evaluation of the damage allowable in structural designs.
As a basic examination for optimizes joint design of CF/PIXA composites using polyimide
resin PIXA with high fracture toughness in this study, a static tensile tests for a double lap
bolted joint were performed. In addition, the experimental results of CF/Epoxy composites
were also reported to compare with these results. During tests, in order to clarify the
relationship between damage progress behavior and bearing strength, the fracture process
under the load-deformation response was monitored using the AE (Acoustic emission)
method. Some of the specimens were unloaded at different loading levels, in order to assess
the damage progress and failure mechanisms using soft X-rays and SEM (Scanning Electron
Microscope) observation of the internal damage. The accumulation of basic data for strength
prediction on the bearing failure of bolted joints was attempted, whilst the mechanism of
bearing failure was examined.
EXPERIMENTAL DETAILS
Material and specimen preparation
In the present work, two types of material systems were selected for comparison:
CF(IM-7)/PIXA (Mitsui Chemistry Co. Ltd.) and CF(IM600)/Epoxy (Q133) (Toho Rayon Co.
Ltd.) materials [10]. These composites were a quasi-isotropic [45/0/-45/90]2S laminate with
same layups, which was made using hot pressing of unidirectional prepreg for 16 layers.
The geometrical configuration and testing setup of the joint specimens has been shown in Fig.
1. The double lap joining configurations were adopted so that the bolt fastener attached the
CFRP specimen to a Chromium Molybdenum steel fixture (with lateral constrains). The
tightening torque was adjusted at 12 ～ 13kgf-cm. All specimens were made with
width-to-diameter ratio (W/d=8), edge-distance-to-diameter ratio (e/d=3).
Experimental procedures
To establish a simpler experimental technique for the bearing strength evaluation of bolted
P
38.1 (w)
Target
Specimen
CFRP specimen
4.8 (d)
AE sensor
15 (e)
Target
Bolt
ZIMMER
camera
INSTRON
test machine
Metal fixture
Bolt
HP A/D
converter
Loading steel plate
Clamping head
Thickness, t=2.24
UNIT: mm
P
Fig. 1 Geometrical configuration and
testing setup of the joint specimens.
PC computeri zed
capture sysytem
Control uni t 101B
Fig. 2 Schematic of measurement system by
non-contacting Electro-Optical Extensometer.
composite joints, a simple and highly functional measurement system has been developed
using a non-contact Electro-Optical Extensometer, which does not require special fixtures,
and has been proposed as a practical test method [11,12]. In this study, the tests were carried
out to evaluate the bearing strength of CFRP composite specimens, using a non-contacting
electro-optical extensometer (ZIMMER Co. product MODEL-100B). The measurement
system configuration is illustrated in Fig. 2. ZIMMER electro-optical extensometer converts
the linear motion of a black-and-white edge (target) into a voltage proportional to
displacement; that is to say, the optical image is converted into an electronic image using a
special photo cathode. In the conventional contact form measuring method, metal fittings
(yokes) have to be installed in which linked both side edges of the specimen with the pinhead
[13,14]. In the present method, measurement of the pin relative displacement is possible only
by bonding the black-and-white target.
Tensile loading was applied using an INSTRON-8501 testing machine; the cross-head speed
was 1.0 mm/min. During static tests, the load-displacement response and acoustic emission
were recorded in order to understand the fracture process of joint specimens. A two-channel
MISTRAS 2001 system (Physical and Acoustic Co.) was used in the AE test, and a sensor
using resonant frequency type with 150kHz (PAC R15I) and the threshold was 55dB. Some of
the specimens were unloaded at different loading levels, and the damage behavior was
observed using soft X-rays and SEM.
1500
16
1250
Load, P (KN)
Bearing load
at P
1000
Ultimate failure
strength
4%d
750
8
P
CF/PIXA
CF/Epoxy
[45/0/-45/90]
4
500
2S
250
d=4.8 mm, e/d=3,
W/d=8, t=2.24mm
P
Bearing strength (MPa)
12
0
0
0
0.5
4%d
1
1.5
2
2.5
3
Displacement,  (mm)
Fig. 3 Typical load-displacement curve of CF/PIXA and CF/Epoxy specimens.
B-S-P6(MPa)
Table 1. Experimental data of bearing strength obtained from each
joint specimen.
B-S-E07a(MPa)
Specimens
CF/PIXA
P-S/N-03
P-S/N-04
P-S/N-07
P-S/N-08
P-S/N-09
Average
CF/Epoxy
E-S/N-04
E-S/N-06
E-S/N-07
Average
4%d Displacement
(KN)
(MPa)
Ultimate Failure
(KN)
(MPa)
Failure Mode
7.78
9.02
8.94
9.71
9.14
8.9
727.7
842.4
835.1
906.6
853.5
835
13.81
14.2
14.07
14.86
14.54
14.3
1289.1
1326.0
1314.2
1381.1
1357.1
1326
Bearing
Bearing
Bearing
Bearing
Bearing
8.63
8.8
8.49
8.6
809.67
825.62
796.46
808
15.85
15.9
15.71
15.82
1486.72
1491.74
1473.45
1485
Bearing
Bearing
Bearing
RESULTS AND DISCUSSION
Bearing strength
A typical load-displacement curve (P-δcurve) of CF/PIXA and CF/Epoxy specimens is
shown in Fig. 3. The bearing strength was obtained based on the bearing load at which the pin
relative displacement was deformed by 4% of the pin diameter (4%D), and the ultimate
failure load from the peak point of the load. An experimental data of bearing strength obtained
from each specimen is shown in Table 1. It can be seen that the average bearing strength of
CF/PIXA specimens is bigger at 4%D, but it is lower at ultimate failure strength.
Bearing failures
AE measurement results [15]
To characterize the internal damage of the joint specimens, the result of monitoring by AE
measurement is shown in Fig. 4. Two materials, it was more clearly shown in this case that
damage behavior was associated with a load increase, with the AE count rate indicating the
AE signal activity. When the applied load was increased to nearly the Knee point (about
10KN), a sharp change in the AE signal was observed. Significant nonlinear behavior was
present on the load-displacement curve at the same time. At this stage, the microscopic
damage of CF/Epoxy specimen was detected near bearing load of 4%D, and it can be seen
that the tendency was easily deformed for compared with CF/PIXA specimen. Following this,
the load response gradually was decreased. Furthermore, these progresses behavior appeared
16.0
16.0
2500
2500
Applied load
1500
AE Count
8.0
1000
4.0
2000
12.0
First peak point
P
4%D
Knee point
AE Count
8.0
1500
1000
4.0
500
500
0.0
0
50
100
150
0
250
200
AE event count
First peak point
Knee point
AE event count
Load, P (KN)
2000
4%d
Load, P (KN)
P
12.0
Applied load
0.0
0
50
100
150
0
250
200
Time (sec)
Time (sec)
(b) CF/Epoxy
(a) CF/PIXA
Fig. 4 Comparison of loading response and AE event count vs. time
for CF/PIXA and CF/Epoxy joint specimens.
1000
1000
CF/PIXA
CF/Epoxy
CF/PIXA
CF/Epoxy
800
AE event count
AE event count
800
600
400
600
400
200
200
0
0
60
70
80
90
100
Amplitude (dB)
(a) Applied Load P=10-12 KN
110
60
70
80
90
100
110
Amplitude (dB)
(b) Applied Load P=12-15 KN
Fig. 5 Distributions of AE event count vs. amplitudes for load levels at 10-12KN and 12-15KN.
16
50
Applied load
P
40
Load, P (KN)
12
Structural
fracture
30
P
4%d
8
-- Local
fracture
4
20
-- Damage
growth
10
CF/Epoxy
CF/PIXA
-- Damage
onset
0
0
50
100
150
200
Accumulative damage (%)
max.
0
250
Time (sec)
Fig. 6 Accumulative damage (AE energy) for CF/PIXA and CF/Epoxy joint specimens.
at stage-by-stage as it can be clearly
seen from the AE signal. It was considered that this
PARA1
PARA1
corresponded to fiber-matrix splitting
and delamination, together with out-of-plane shear
PARA1
cracking.
Fig. 5 shows AE amplitude distributions of each specimen in various loading stages. Both
materials in loading range 10～12KN and 12～15KN, CF/PIXA specimen is manifest at the
low or middle amplitude of 80dB, CF/Epoxy specimen at the high amplitude over 90dB is
comparatively controlled. Based on the observation of SEM photographs after tested, it can be
seen that the fiber micro buckling or matrix crack are correspondent to the AE signal of low
or middle amplitude, and delamination or out-of-plane shear crack are correspondent to the
AE signal of high amplitude.
CF/PIAX
16
CF/Epoxy
Stage Ult.
P Max (100%)
Stage...
Applied load, P (KN)
12
Stage 3
(75%)
Stage 2
(60%)
P4%d
8
Stage 1
(50%)
4
0
0
0.5
1.0
1.5
2.0
Displacement, δ (mm)
Fig. 7 X-radiographs for bearing damage under the various load conditions.
The variation of accumulated damage rate (accumulated energy) that shows the total of
damage production in the same load level is shown in Fig. 6. The AE signal of CF/PIXA and
CF/Epoxy specimen was detected in 9KN and 7.8KN respectively. CF/Epoxy was lower than
CF/PIXA. And it was observed that when the applied load was increased to 10～12KN range,
the accumulated damage rate began to change in AE signal. Until the final fracture, the
difference of accumulated damage rate is about 15% for two specimens. It was also accorded
to the results of ultimate failure strength in Fig. 3, indicated that the fracture toughness of
PIXA resin is high.
X-radiographs results
Fig. 7 shows X-radiographs for the bearing damage under various loading conditions.
Corresponding with the AE results, it is possible to evaluate the damage state and scale of the
damage production. In the initiation stage (50% max. load) prior to reaching the P4%D load, it
can be seen that microscopic damage has occurred and this agrees well with the AE results.
But this damage does not affect very much the response of load-displacement curve. It was
also proven that the damage scale of CF/PIXA specimen is small. Upon increasing the load,
the extent of damage became more extensive around the hole, significant nonlinear behavior
was present on the load-displacement curve, and this tendency continued until reaching the
final fracture. The damage scale of CF/Epoxy specimen is relatively large.
SEM photographs results
During the tests, in order to observe the internal damage of bearing some of the specimens
1
○
2
○
1
○
2
○
Kink band
band of
of 00oo plies
plies
Kink
Washer
o
Kink band of 0 plies
CF/PIXA (Load: 60%Pmax)
1
○
2
○
1
○
2
○
Kink band of 0o plies
Washer
CF/Epoxy (Load: 60%Pmax)
Kink band of 0o plies
Fig. 8 SEM photographs of the bearing damage at 60% of maximum load.
Washer
Washer
CF/Epoxy (Load: 60%Pmax)
CF/Epoxy (Load: 75%Pmax)
Fig. 9 SEM photographs of the bearing damage at 75% of maximum load.
CF/PIXA
(Load: 100%Pmax)
CF/Epoxy
(Load: 100%Pmax)
Fig. 10 SEM photographs of the bearing damage at 100% of maximum load.
were unloaded at different load levels, and the tested specimens were cut along the centerline
hole. Micrographs were made of these cross-sections using SEM. Fig. 8 to Fig. 10 shows
SEM photographs of the bearing damage along the loading direction, loaded to 60%, 75%,
and 100% of the maximum load.
At 60% of the maximum load, the unusual kink band damage in the compressive failure can
be clearly seen in the 0゜plies, whilst local matrix cracks and delamination also occurred
under the washer area, CF/Epoxy specimen was more clearly. Next, the load was increased to
75% of the maximum load, with out-of-plane shear cracks occurring more frequently than the
Crushing
damage
Contact
surface
(a)
Washer --
Washer --
Loss of
washer
support
Fixture --
(b)
(c)
Fig. 11 Schematic description of the bearing damage mechanisms in bolted joint specimens,
(a) damage at contact surface, (b) damage within washer area, (c) damage under washer area.
previous load level, and fiber-matrix splitting cracks can also be found in the 0゜plies. Such
shear cracks that CF/PIXA occurred in the external surface layer, CF/Epoxy occur in full
section of laminate, and the damage scale is also bigger than CF/PIXA. These damage
patterns continued until reaching the final fracture, which can be understood from the case of
the maximum load. In other words, it is conceivable that the load is reduced stage-by-stage on
the P-δcurve due to the shear cracks which occurred intermittently.
Summarization of damage pattern
Based on the above SEM photograph observations, a schematic description of the damage
pattern is shown in Fig. 11. The damage pattern features of bearing failure can be explained as
following.
(A) Damage at contact surface; it is crushing damage that occurred in contact surface
between bolt/hole. This damage progress will be depending on the response of
material stiffness and friction effect.
(B) Damage within washer area; it is accumulated damage that occurred in washer area,
and formation and accumulation of the damage are dependent on the clamping
pressure. Since the out-of-plane deformation of laminate was suppressed by
clamping pressure the compression failure behavior was mainly expanded along the
in-plane direction, and the damage progress has been presented at stage-by-stage.
(C) Damage under washer area; the damage such as the large-scale splitting cracks
occurred under washer area due to the damage gradually reached the saturation state
in washer area, and the joint response rapidly lowers.
Based on these experimental observations, it can be concluded that the kink band,
delamination, fiber-matrix shearing, and fiber-matrix splitting failure formed the major failure
modes of the shear cracks. For the growth of shear cracks it can be concluded that the
influence of the washer area and bolt clamping conditions is large.
CONCLUSIONS
In order to evaluate the bearing strength and damage behavior of the joint specimens, an
experimental investigation of CF/PIXA and CF/Epoxy bolted composite joints was performed,
and the relationship between load-displacement curve, AE characteristic and failure
mechanism was also examined. There are correlations of the damage pattern and damage
progress that the activity of the AE signal corresponds to the joint response. Furthermore,
corresponding with results observed using X-rays and SEM photography, the bearing failure
mechanisms were explained. Based on these, the following remarks can be made:
1. The kink band of 0゜plies and delamination appeared to be the dominant mode in
the onset of damage. Out-of-plane shear cracks and fiber-matrix splitting cracks
were the major reason for final failure.
2. Lateral constraint condition suppressed out-of-plane deformation, and it was
recognized as an important parameter to improve the joint strength.
3. To evaluate the lateral clamping effect on the restrain of out-of-plane deformation,
the construction of a 3-D damage model and development of analysis tools should
be become a focus of further investigation.
REFERENCES
1. Hart-Smith, L. J. “Bolted Joints in Graphite/Epoxy Composites”, NASA Langley Report
CR-144899, 1976.
2. Matthews, F. L. “Joining Fiber-Reinforced Plastics”, New York: Elsevier Applied
Science (1986).
3. AGARD Conference Proceedings CP-427, “Behaviour and analysis of mechanically
fastened joints in composite structures”, North Atlantic Treaty Organization (1988).
4. AGARD Conference Proceedings CP-590, “Bolted/ Bonded Joints in Polymeric
Composites”, North Atlantic Treaty Organization (1998).
5. Xiao, Y., Wang, W. X., Takao, Y. and Ishikawa, T. “The Effective Friction Coefficient of
a Laminate Composite, and Analysis of Pin-Loaded Plates”, Journal of Composite
Materials, Vol. 34, No. 1 (2000), pp. 69-87.
6. Eriksson, I. 1990. “On the Bearing Strength of Bolted Graphite/Epoxy Laminates,”
Journal of Composite Materials, 24:1264-1269.
7. Wang, H. S., C. L. Hung and F. K. Chang. 1996. “Bearing Failure of Bolted Composite
joints Part I: Experiments,” Journal of Composite Materials, 30(12): 1284-1313.
8. Wu, P. S. and Sun, C. T. April 1997. “Bearing failure in pin contact of composite
laminates,” Proceedings of the 38th AIAA/ASME/ASCE/AHS/ASC Structures, Structural
Dynamics and Materials Conference, pp.1974-1983.
9. Camanho, P. P., Bowron, S. and Matthews, F. L. “Failure mechanisms in bolted CFRP”,
Journal of Reinforced Plastics and Composites, Vol. 17, 1998, pp.205-233.
10. Ishikawa, T. and Matsushima, M. Proceedings of 11th ICCM, Gold Cost, Australia, July
1997, Vol.II, pp. 1-9.
11. Xiao, Y., Hayashi, Y. and Ishikawa, T. “Consideration on Bearing Strength Evaluation
Method in Composite Materials by Electro-Optical Extensometer”, April 1999.
Proceedings of Annual Meeting of Japan Society for Composite Materials, pp.77-78. (in
Japanese).
12. Xiao, Y. and Ishikawa, T. “Non-contact Measurement for Bearing Strength of
Mechanically fastened Joints in CFRP Composites”, Journal of the Japan Society of
Composite Materials, 26(6): pp.213-218. (in Japanese).
13. JIS K 7080, 1991. “Testing methods for bearing strength of carbon fiber reinforced
plastics”, Japanese Industrial Standard.
14. ASTM D 953-95, “Standard Test Method for Bearing Strength of Plastics”, American
Society for Testing and Materials (1995).
15. Xiao, Y., Hayashi, Y. and Ishikawa, T. “Examination of Acoustic Emission on Failure
Processes of Bolted Composite Joints”, 2000-10. Proceedings of The 25th Symposium
For Composite Materials (Japan-Korea Joint Session), pp.77-78.