Optimum Bolted Joints For Hybrid Composite Materials

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Composire Sbucwes Vol. 38, No. 1-4, pp.

329-341, 1997
0 1997 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0263-8223/97/$17&l + 0.00
ELSEVIER
PII:SO263-8223(97)00067-6

Optimum bolted joints for hybrid composite


materials
Je Hoon Oh, Young Goo Kim & Dai Gil Lee*
Mechanical Design Labomtory withAdvanced Materials, Department of Mechanical Engineering, ME3221 Korea Advanced
Institute of Science and Technology, Gusong-dong, Yusong-gu, Taejon-shi, Korea 305 701

The optimum bolted joints for hybrid composite materials composed of


glass-epoxy and carbon-epoxy under tensile loading were investigated. The
design parameters considered for the bolted joints were ply angle, stacking
sequence, the ratio of glass-epoxy to carbon-epoxy, the outer diameters of
washers and the clamping pressure. As bearing failure was desirable for
bolted joints, the geometry of the bolted joint specimen was designed to
undergo bearing failure only.
By inspecting the fracture surfaces of the specimens it was found that
delamination on the loaded periphery of the holes and extensive damage on
the edge region constrained by a washer occurred. To assess the
delamination of the hybrid composite materials, three-dimensional stress
analysis of the bolted joint was performed using a commercial finite-
element software and compared with the experimental results. 0 1997
Elsevier Science Ltd.

INTRODUCTION because it usually becomes the weakest part of


the structure [6,7]. There are two types of
The fiber-reinforced composites have been used joints: mechanical and adhesively bonded joints.
widely in aircraft and spacecraft structures, as The mechanical joint has several advantages
well as in sports and leisure goods, because they over the adhesively bonded joint, which requires
have high specific stiffness, high specific careful surface treatment of the adherend, is
strength, high damping and a low coefficient of affected by service environment and is difficult
thermal expansion. These properties cannot be to dismantle for inspection and repair, although
obtained from conventional metals such as steel it does distribute load over a larger area than
and aluminum [1,2]. In addition, the structures the mechanical joint. However, the mechanical
made of composites have a high natural fre- joint requires holes to be drilled for bolts and
quency as well as a light weight, therefore they rivets, which causes unavoidable stress concen-
are increasingly employed in rotating structures trations as well as a weight penalty due to the
such as power transmission shafts of aircrafts bolts and rivets [8].
and automotives [3], in machine elements such Three different types of fasteners, such as
as machine tool spindles [4] and robot arms [5]. self-tapping screws, rivets and bolts, are usually
When composites are employed as structural used in mechanical joints. Among these fasten-
materials, joining of composites to other ers, the bolted joint was found to be most
materials is necessary because manufacturing efficient for the mechanical fastening of compo-
the whole structure using only composites is not sites [9]. However, the analysis of mechanical
generally feasible. Therefore, the joint used in joints has not been very satisfactory because of
composite structures, with only a few excep- the friction between bolts and holes, the
tions, determines the structural efficiency material non-linearity and the presence of
three-dimensional stresses and strains. Thus,
*Author to whom correspondence should be addressed. many investigators have experimentally studied

329
330 J. H. Oh, Y G. Kim, D. G. Lee

the behavior of mechanical joints of CFRP and composites such as carbon-epoxy and glass-
GFRP. epoxy. Therefore, there are few data available
Collings [9] investigated the effect of stacking on the joints for hybrid composites [26].
angle on the bearing strength with respect to In this work, the optimum bolted joints for
three different composites. Eriksson [lo] hybrid composite materials composed of glass-
showed that bearing strength was strongly epoxy and carbon-epoxy under tensile loading
affected by a lateral constraint of the material were investigated. The geometry of the bolted
surrounding the loaded hole. Quinn & Mat- joint specimen was designed to induce bearing
thews [ll] reported a significant effect of the failure because bearing failure is preferable for
stacking sequence on the bearing strength of the bolted joint. The design parameters were
GFRP. Herrington & Sabbaghian [12] studied ply angle, stacking sequence, ratio of glass-
the effect of radial clearance between the bolt epoxy to carbon-epoxy, outer diameter of the
and washer on the bearing strength of the com- washer and clamping pressure.
posite bolted joints and showed that bearing The optimum ply angle for hybrid composites
strength increased as the radial clearance was determined using the static test results for
decreased. They also found that the ratios of the bolted joints of carbon-epoxy and glass-
width-to-diameter (w/d) and edge distance-to- epoxy composites with respect to ply angle.
diameter (e/d) influenced the failure mode, and Then the joint specimens for the hybrid compo-
the bearing strength was increased and sites were tested under tensile loading with
approached a limiting value as the ratios were different stacking sequences and ply number
increased [13-151. Kim & Whitney [16] found ratios of glass-epoxy to carbon-epoxy compo-
that elevated temperature and increased mois- site. The effects of the outer diameter of the
ture content of the laminate reduced the washer and the clamping pressure of the bolt on
bearing strength. Akay [17] studied the static the bearing strength were also investigated.
and dynamic bearing strengths of uni-direc- In order to investigate the delamination of
tional and woven carbon-epoxy composite the hybrid composite materials three-dimen-
when test pieces were hygrothermally treated. sional stress analyses for the bolted joint were
Most models for the strength prediction of performed using a commercially available finite-
bolted composite laminates adopted elastic two- element software and then compared with the
dimensional plane-stress analysis using a experimental results.
finite-element method [l&21]. Chen et al. [22]
used the incremental restricted variational prin-
ciple and the transformation matrix, derived EXPERIMENTS
from three-dimensional contact kinematic con-
ditions, to carry out a three-dimensional contact Because the failure of bolted joints should not
stress analysis of a composite laminate with bol- occur catastrophically, in this work the speci-
ted joints. Graham et al. [23] performed a men geometry was designed to induce bearing
two-dimensional contact stress analysis assum- failure. For this purpose, both the ratio of
ing that bolts and washers were rigid and that width-to-diameter and of edge distance-to-
bolts had a perfect fit. Smith et al. [24] estab- diameter were set at 5. The geometry of the
lished a simple three-dimensional model to bolted joint is shown in Fig. 1.
predict the strength of bolted composite lami- In order to determine the ply angle of the
nates, and studied the effects of the coefficient hybrid composites the static tests of the bolted
of friction between washers and laminates. joints for carbon-epoxy and glass-epoxy were
Although hybrid composites composed of car- performed with respect to the ply angle. The
bon-epoxy and glass-epoxy are in widespread number of plies for both the carbon-epoxy and
use in automotive and aerospace propeller the glass-epoxy composite were 20. The stack-
shafts, leaf springs and fly wheels due to their ing sequences for both the composites were
improved mechanical behavior, such as their ]0,/(0/90),/90,1, and [02/( + f&/90& (0 = 15, 30
high specific stiffness, high specific strength, and 45”).
high impact energy, cost reduction, good corro- The carbon-epoxy composite and the glass-
sion resistance and easy fabrication [25], most epoxy prepregs used were USN150 and
studies have been focused on the behavior of UGN150, respectively, both manufactured by
mechanically fastened joints for non-hybrid Sunkyung Industry (Suwon, Korea). Table 1
Optimum bolted joints for hybrid composites 331

Steel

Washer (thickness = 2mm)

w/d = 5 e/d = 5
Number of plS@r= 20 plier
Specimen
Fig. 1. Geometry of the specimen.
Tab
shows the uni-directional mechanical properties
Fig. 3. Schematic diagram of loading fixture.
of the composite materials and Fig. 2 shows the
curing cycle for the composite materials.
A glass fabric-epoxy composite was used for temperature with an Instron 4206, a static
the tabs on the composite specimens, and the material testing machine with a test speed of
tensile tests were performed using the loading 1 mm/min. Five specimens were tested for each
fixture illustrated in Fig. 3. The 8 mm bolt with point.
class 10.9 was used for the bolted specimens. Typical load-displacement curves are shown
Washers, 2 mm thick, with the same inner diam- in Fig. 4, where the first peak loads (A) indicate
eter as the bolt and a 20 mm outer diameter the local delaminations around the holes under
were used. The specimens, clamped using fin- the washers. The damage development in the
ger-tightened bolts, were tested at room vicinity of the hole boundary at the peak point
(A) was measured using an ultrasonic C-scan as
able 1. Material properties of the carbon-epoxy shown in Fig. 5. From the result of the C-scan,
(USNlSO) and glass-epoxy (UGNlSO) composites it was found that damage developed along the
Carbon-epoxy Glass-epoxy loaded half of the hole circumference. The load
(USNISO) (UGNlSO) was continuously increased because the swelling
EL VW 130.0 43.5 of the composite in the thickness direction, due
to the Poisson effect, induced the clamping
ET
% (GPa)
(GPa) 8.0
6.0 :.:
vLT 0.28 0.25
-T (MW 1800 1000
Y (MPa) 60 50

S
Ply(MPa)
thickness (mm) 75
0.15 0.25

A Carbon/Epoxy

J
0 1 2 3 4 5 6
Displacement (mm)

Fig. 4. A typical load-displacement curve for the carbon-


Fig. 2. Cure cycle for composite materials. epoxy bolted joint and the glass-epoxy bolted joint.
332 J. H. Oh, Y G. Kim, D. G. Lee

Table 2. Stacking sequences of the carbon-epoxy and


glass-epoxy composites
Material Stacking sequence

Carbon-epoxy Pdw9w9021s 69.3


PA f 15)J902ls 95.8
Pd( * 3W90,ls 69.8
W( f 4W90,ls 46.1
Glass-epoxy Pdw9w90*1s 24.4
W( + w3/9021s 33.1
[W f 3W90,ls 26.1
YW rt 4W90,ls 19.2
‘Ek, Laminate longitudinal modulus.

390 400 410 420 430 440 450 460 470 occurred when the stacking sequence was
0.19mm X SCALE/mm P,"OTHEIGHT= 83 mm [OJ( _+8),/90,1s (6’= 15, 30 and 45”). Figure 6a
Fig. 5. Results of the ultrasonic C-scan of the bolted joint. shows the bearing strength and failure modes of
carbon-epoxy with respect to the ply angle. On
inspecting the fracture surfaces of the speci-
pressure in the region of the laminate con- mens, delamination of the carbon-epoxy
strained by the washer, which delayed the specimen on the loaded side of the holes and
delamination of the specimens. Therefore, in extensive damage at the washer edge outside
this work, the ultimate failure load rather than the constrained region were found. From the
the first peak load was taken as the joint test results, it was found that delaminations
strength load because the joint was not frac- were more critical to the joint failure than in-
tured at the first peak load (A) and the rate of plane damage.
the increase of the load was not decreased. In order to investigate the effect of a 90” ply
The ultimate bearing strength, g,,, was on the joint strength, the specimen with a stack-
defined by the following equation ing sequence of [03/( +45),/+451s was tested
and the result was compared with that of
P [OJ( +45)J90,],. The bearing strengths of the
cTb= - specimens with [O,/( &-45),/+45], and
dt were 859 and 880 MPa, respec-
[O*/(* 45)3/90&
tively, where the bearing strength was higher
where P is the ultimate failure load, d is the when a 90” ply was contained in the laminates.
hole diameter and t is the thickness of the lami- From the static test results of a glass-epoxy
nate. composite with a stacking sequence of
[02/(0/90),/9021s and [O,/( + 8),/90& (0 = 15, 30
Static tests of joints for the carbon-epoxy and and 45’7, it was found that the bearing strengths
glass-epoxy composites of the two specimens were similar. The speci-
men with [02/(0/90),/90,], failed in the
In order to determine the optimum ply angle combined modes of bearing and shear-out fail-
for hybrid composites, the bolted joints for the ures. The specimen with [OJ( f 15)3/90,]s was
carbon-epoxy and glass-epoxy composites were failed in the combined modes of bearing, ten-
first tested with respect to the ply angle. Table 2 sion and shear-out failures, while the specimens
shows the stacking sequences for the carbon- with a [O,/( & 0),/90,], (0 = 30” and 45”) joint
epoxy and glass-epoxy composites. From the failed in bearing failure mode. Figure 6b shows
static test results of the carbon-epoxy compo- the bearing strength and failure modes of the
site with a stacking sequence of [0,/(0/9O)J90,], glass-epoxy composite with respect to the
and [OJ( + 13)~/90~]~(0 = 15, 30 and 457, it was stacking sequence.
found that the bearing strength was increased as Similar to the case of the carbon-epoxy com-
the ply angle 0 was increased. When the stack- posite, delamination of the glass-epoxy
ing sequence was [O&Y90),/90,],, the specimen on the loaded side of the holes and
combined modes of tensile and shear-out extensive damage at the washer edge outside
failure occurred, and the bearing failure mode the constrained region were found. During the
Optimum bolted joints for hybrid composites 333

tests it was found that the area of the specimen bon-epoxy with [OJ( + 45)J90,], yielded the
between the hole and the edge was deformed highest bearing strength and the glass-epoxy
considerably out of the laminate plane due to with [02/( +30)$X& and [Od( f 45)J90,1s
the low modulus of the glass-epoxy composite. yielded similar high bearing strengths. There-
This out-of-plane buckling mode was also fore, for the hybrid composites, the stacking
observed by Kretsis & Matthews [27]. sequence was [OC/ + 45G/ + 45C/9OC], (C: car-
bon-epoxy, G: glass-epoxy) containing carbon
Static tests of the joints for the hybrid plies in the O”, 90” and +45” directions and
composites glass plies in the +45” direction.
In order to investigate the effects both of the
From the static tests of the carbon-epoxy and clamping pressure of the bolt and the outer
glass-epoxy composites, it was found that car- diameter of the washer on the bearing strength,
the stacking sequence and the ratio of carbon-
epoxy to glass-epoxy which gives the highest
bearing strength was used in the test of speci-
mens. As the thickness of carbon-epoxy was
different to that of glass-epoxy, as shown in
Table 1, the percentage of ply number is dif-
ferent to the volume fraction. Therefore, the
volume fractions of the glass-epoxy with ply
* 600 -
number percentages of 20, 40 and 60 are 17.2,
4

2
a, 500 -
35.8 and 55.6%, respectively. Table 3 shows the
s
r/l
stacking sequences of the hybrid composites
w 400 -
d
which have the stacking sequence type [OC/
‘i:
$ 300 -
rt:45Gl f 45C/9OC],. The specimens with 14
m different stacking sequences were tested when
200 _I I the volume fractions of glass-epoxy were 17.2
and 35.8%, while the specimens with seven dif-
ferent stacking sequences were tested when the
I I
0 I I volume fraction of glass-epoxy was 55.6%.
o/so 15 30 45

Stacking Angle (deg)


When the volume fraction of the glass-epoxy
was 17.2% (OC: 20.7%, f 45G: 17.2%, + 45C:
(4
1000
41.4%, 9OC: 20.7%), type G yielded the highest
bearing strength (941 MPa) and type H yielded
000 the lowest bearing strength (834 MPa), the dif-
800
ference being 107 MPa as shown in Fig. 7.
From Fig. 7a and b, it was found that the
2 700 bearing strength was dependent on the location
r

600
of the + 45G and f 45C plies. The difference in
5
bearing strengths between types A-G compo-
2
al 500 sites and types H-N composites was about
s
cn
M 400
60 MPa, although the only difference in the two
d
.d types of composites was the location of the
L
t 300 &45G plies. Therefore, it was concluded that
q

200
the location of the &45G plies was important
for the bolted joint in composite materials.
100 When the 9OC plies were located in the
vicinity of the k45G plies (type F) and the
0

o/so 15 30 45
+45C plies were located in the vicinity of the
Stacking Angle (deg) +45G plies (type M), the bearing strength of
0-J)
type F was higher than that of type M. Simi-
larly, type B as shown in Fig. 8a, where the
Fig. 6. Joint bearing strengths of the CFRP and GFRP
with respect to the stacking angle of [Od( + 0)J90,1s. (a) +45G plies are located between the OC and the
CFRP. (b) GFRP. +45C plies, had a lower bearing strength than
334 J. H. Oh, Y G. Kim, D. G. Lee

that of type I as shown in Fig. 8b, where the the hybrid composites. As shown in Fig. 11, the
_+45G plies were located between the &-90C bearing strength decreased for all the stacking
and the k45C plies. Therefore, it was con- sequences as the volume fraction of glass-epoxy
cluded that the interlaminar stress was increased. In case of type A, the bearing
concentration was decreased due to the smaller strengths of joints for the pure carbon-epoxy
difference in stiffness of each ply when the and the pure glass-epoxy are also presented in
+45G plies were located between the 90C and Fig. 11. Although glass-epoxy is cheaper and
the +45C plies. more impact-resistant than carbon-epoxy, it is
When the volume fraction of the glass-epoxy important to determine an appropriate volume
was 35.8% (OC: 21.4%, f45G: 35.8%, +45C: fraction of glass-epoxy as the bearing strength
21.4%, 90C: 21.4%) type G yielded the highest of the joint decreases as the volume fraction of
bearing strength (917 MPa) and type F yielded glass-epoxy increases.
the lowest bearing strength (792 MPa), as When the stacking sequence of the hybrid
shown in Fig. 9. composite was [OG/ & 45C/90G],, where the
When the volume fraction of the glass-epoxy volume fractions for glass-epoxy and carbon-
was 55.6% (OC: 22.2%, +45G: 55.6%, +45C: epoxy were 35.8 and 64.2%, respectively (OG:
O%, 90C: 22.2%), type G yielded the highest 17.9%, f45C: 64.2%, 90G: 17.9%), the speci-
bearing strength (788 MPa) and type C yielded mens with seven different stacking sequences,
the lowest bearing strength (695 MPa), as as shown in Table 4, were tested.
shown in Fig. 10. In this case it was found that the bearing
Figure 11 shows the bearing strength with strength increased as the volume fraction of the
respect to the volume fraction of glass-epoxy in +45C plies was increased, which is a similar

Table 3. Stacking sequences of the hybrid composites [OC/+ 45G/ + 45C/9OCl s


Stacking pattern’ Type Stacking sequence
[OC/ f 45Glf 45C/9OC],OC: A [(OC)d( -t 45C),/ +45G/(9OC),],
20.7% + 45G: 17.2% f 45C: B [(9OC)d( f 45C)d +45G/(OC)zls
41.4%9OC: 20.7% C [( + 45C)d +45G/(OC)d(9OC),l,
D [(9OC)d(OC),/( I!Z45C),/ f 45G],
E [(OC)d(9OC)d( f45C),/ f 45G],
[( ~45C)d(OC)d(9OC)d~ 45G],
: [+45C/-45C/+45C/(OC)~-45C/+45G/(90C)&
H [(OC)d + 45G/( I!I45C)J(9OC),ls
I [(90C)d+45G/( +45C),/(OC)z],
J [ f 45G/( + 45C),/(OC)d(9OC),Is
K [(90C),/(OC)d +45G/( + 45C),],
L [(OC)d(9OC),/~ 45G/( * 45C&
M [ f 45G/ -t 45C/(OC)d(90C),/ + 45C]s
N [+45G/-45G/+45C/(OC),/-45C/+45C/-45C/(90C),],
[OC/ &-45G/ + 45C/9OC],OC: A [(OC)z/ +45C/( + 45G)d(90C),],
21.4% f 45G: 35.8% &-45C: B [(9OC)J+45C/( f 45G)d(OC),]s
21.4%90(7: 21.4% C [-t 45C/( -t 45G)d(OC)d(9OC),l,
D [(9OC),/(OC),/ f 45C/( + 45G),],
E [(OC),/(9OC)Jk45C/( f45G)z],
[ + 45C/+ 45G/(OC)d(90C)d I+45G],
: [+45C/-45C/+45Gl(OC),i-45G/+45Gl-45G/(90C),],
H [(OC),/( z!z45G)d +45C/(9OC),],
I [(9OC),/( f 45G)df45C/(OC),],
J [( I!Y45G),/ rt 45C/(OC)d(9OC),]s
K [(9OC)d(OC),/( I!T45G),/ &45C],
[(OC)d(9OC)d( f45G),/ zh45C],
k [( I!I45G)d(OC)d(9OC)d( f 45C)ls
N [+45G/-45G/+45G/(OC),/-45G/+45C/-45C/(90C),],
[OC/ + 45G/ + 45C/90C]sOC: A [(OC)dt + 45G)d9OC),ls
22.2% + 45G: 55.6% f 45C: B [(9OC)d( f 45G)dOC),l,
0%9OC:22.2% C [( !I 45G)~(OC)~(9OC)&
D [(9OC)d(OC)d( f 45G),l,
E [(OC)d(9OC)d( + 45G),ls
F [( * 45G)J(OC)d(9OC)J ?I 45G],
G [+45G/-45G/+45G/(OC)~-45G/+45G/-45G/(90C)~]s
‘C, Carbon-epoxy; G, glass-epoxy.
Optimum bolted joints for hybrid composites 335

trend to that found for [OG/ + 45G/ + 45Cl epoxy, type G of [OC/+ 3OG/If:45C/9OC], was
90G]s. Therefore, the specimens with also tested and compared to [OC/&45G/f 45C/
[OG/ + 45C/90G], (OG: 17.9%, +4X: 64.2%, 9OC]s. As shown in Fig. 13, when the volume
90G: 17.9%), which had a high of volume frac- fractions of the glass-epoxy were 17.2 and
tion of +45C plies, were tested. Type G yielded 35.8%, the bearing strength of the joint contain-
the highest bearing strength (974 MPa) and ing &30G was a little higher than that of the
type C yielded the lowest bearing strength *45G, while the bearing strength of the joint
(839 MPa), as shown in Fig. 12. containing +30G was a little lower than that of
As it was found that the bearing strength of the &45G when the volume fraction of the
[Od( f 45),/90,3s was similar to that of glass-epoxy was 55.6%.
[OJ( +30)J902]s from the static test of glass- The hybrid composites of type G [OG/+45C/
90G]s (OG: 17.9%, f45C: 64.2%, 90G: 17.9%)
which had the highest bearing strength, were
tested with respect to the clamping pressure of

loo0
t z
the bolt and the outer diameter of the washer,
when the inner diameter of the washer was
same as that of the bolt and the outer diameter
of the washer was 20 mm. After lubricant was
applied to the bolt threads to reduce the varia-
tion of preload, the clamping pressure was given

5
VI
I I

G F D C A B E

Stacking Sequence Type

loo0 4

900 _ (a)
T
3
T

f
4
2 600
0) _
f:
rn

f 45G 14.4
I

f 45c 20.9

N I L K H .I II
Stacking Sequence Type

@I

Fig. 7. Joint bearing strengths of the hybrid joints with


(b)
respect to stacking sequences ([OC/If:45G/k 45C/9OC]s Fig. 8. Effective longitudinal modulus of the plies in the
(OC, 90C: 20.7%, f45G: 17.2%, f45C: 41.4%)). (a) vicinity of the +45G ply. (a) Case of low bearing strength
Type A-type G. (b) Type H-type N. (type B). (b) Case of high bearing strength (type I).
336 J. H. Oh, Y G. Kim, D. G. Lee

1000 1000

900 900
3 3
9 B
c 5
Tlcl F!
B 600 P m 600
5
r/l
5
ul

2
I I
‘i:
:
m. 700

600 I I I I I I
i 600 I I I I I

G F D 6 C E A G F D B A E C
Stacking Sequence Type Stacking Sequence Type

(4 Fig. 10. Joint bearing strengths of the hybrid joints with


respect to stacking sequences ([OC/ f 45G/+ 45C/9OC],
(OC, 90C: 22.2%, +45G: 55.6%, +45C: 0%)).

N K L J M H I

Stacking Sequence Type

@I
0 10 20 30 40 50 60 70 80 so 100
Fig. 9. Joint bearing strengths of the hybrid joints with
Volume Fraction of Glass/Epoxy (X)
respect to stacking sequences ([OC/ + 45G/ + 45C/9OC],
(OC, 90C: 21.4%, +45G: 35.8%, k45C: 21.4%)). (a) Case Fig. 11. Bearing strengths with respect to the volume frac-
of type A-type G. (b) Case of type H-type N. tion of the glass-epoxy and laminate type.

Table 4. Stacking sequences of the hybrid composites [0G/45C/9oGl,


Stacking pattern’ Type Stacking sequence
[OG/ +45C/90Gls A [(OG)d( + 45C)J(9OG),ls
OG: 17.9% ](9OG)dt f 45CMOGMs
+ 45C: 64.2% E ]( * 45CMOG)J(9OG),l,
90G: 17.9% ](9OG)d(OG)J(45C)&
: ](OG),1(9OG)d( * 45CMs
[( f 45C)d(OG)&9OG),l+ 45Cls
E ]+45C/-45C/+45C/(OG)z/-45C/+45C/-45C/(90G),1,
‘C, Carbon-epoxy; G, glass-epoxy.
Optimum bolted joints for hybrid composites 337

and d is the nominal bolt diameter (mm). Then


eqn (2) becomes

T
opress= _ (3)
t Kd(d; - d2)

or
~~~~~~
(MPa) = 2.37T (N-m) (4)
where d,,, is the outer diameter of the washer.
The torque range used was O-40 N-m, which
was O-95 MPa in terms of the axial load.
Figure 14 shows the bearing strength with
respect to clamping pressure. As shown in
600 I/ Fig. 14, the bearing strengths of the hybrid com-
G F A B C D E posites were increased as the clamping pressure
Stacking Sequence Type was increased up to 71.1 MPa, and it then con-
Fig. 12. Joint bearing strengths of the hybrid joints with verged to a constant value which was 1.3 times
respect to stacking sequence ([OG/+45C/90G], (OG, 90G: higher than the value without preload.
17.9%, f 45C: 64.2%)).
The bolted joints for the hybrid composites
were tested by varying the outer diameter of the
by a torque wrench [28]. The torque T (N-m) washer, ranging from 1.5d (12 mm) to 4.5d
required for a given axial preload F(N) is cal- (36 mm) when the inner diameter of the washer
culated as follows [29] and the clamping pressure of the bolt were
8 mm and 23.7 MPa, respectively. As shown in
Fig. 15, as the outer diameter of the washer was
d
T=KF- (2) increased up to 2.5d (20 mm), the bearing
1000 strength was increased and then saturated.
When the outer diameter of the washer was
where K is the torque factor (0.2 on average) smaller than 2.5d (20 mm) bearing failure
occurred; however, tension failure occurred

1000

0 0
0 1200

0
900
3 ~1100
9 B

_o
+a 1000
2
0
5
rf~ 900
0
2
‘C
g 600
q

o [OC/f45G/f45C/90C'$
700
o [ OC/f30G/f45C/90C],

I 1 I I 600 L I , 1 I 1 I I I

10 20 30 40 50 80 0 10 20 30 40 50 60 70 80 90 100
Volume Fraction of Glass/epoxy (9;) Clamping Pressure (MPa)

Fig. 13. Bearing strengths of the [OC/k45G/+45C/9OC]s Fig. 14. Bearing strength with respect to the clamp-up
and [OC/ ~3OG/+45C/9OC]s with respect to the volume pressure for type G of [OG/&-45C/9OG], (washer outer
fraction of the glass-epoxy. (All cases are type G.) diameter = 20 mm).
338 J. H. Oh, Y G. Kim, D. G. Lee

outside edge were found from the joint tests of


*2oo
’ the hybrid composites, delamination of the bol-
ted joints for the hybrid composite materials
1100
was investigated using ANSYS, a commercial
4, ;h 0 m
finite-element software, and compared with the
2
2 1000
experimental results. A cosine load distribution
. was used for the contact force between the
5

I
composites and the bolt. The solid 46 element
2 900
5
(u
of the ANSYS element library, which enables
VI
the assignment of the ply angle, stacking
P

1
2 sequence and ply thickness, was used for the
‘C 600

m
zi stress analysis. Figure 17 shows the finite-ele-
ment model for the bolted joint. Owing to
700 Bearing
Failure --I I- Tension
Failure symmetry, only a quarter of the laminate was
modeled and the finger-tightened washer effect
1

600 I31 was simulated by constraining the z-direction


10 15 20 25 30 35 40 displacements of the elements under the
Washer Outer Diameter (mm) washer.
Fig. 15. Bearing strength with respect to the outer diam- The delamination of the hybrid composites
eter of the washer for type G of [OG/~45C/90G]s was predicted using the Ye-delamination failure
(clamping pressure = 23.7 MPa).

(a)
when the outer diameter was larger than 3d ANSYSSOA
(24 mm) because the gap distance (w’) between 20:06:13
PLOTNo 2

the washer edge and the specimen edge was ELEMENTS


TYPENUM

reduced, as shown in Fig. 16, as the washer size XV=,


YV=I
was increased. zv= 1
DIST=O078639
XF=O.I
YF = 0.0,
ZF = 0.001,
CENTROIDHIDDEN
STRESS ANAYSIS

Because delamination on the loaded side of the


holes and the extensive damage at the washer

SPECIMEN FOR THE BOLTED JOINT

W-

t-
tb)

% 7 ANSYS 5 OA
20:06:01

t-7
PLOTNo I
ELEMENTS
TYPENIJM

T-- t-rdh
c
xv=,

1
Y"=I
zv= I

‘-i-
DIST=0.013302
XF =0.030526
YF =0.009524
ZF=O.OOll
CENTROID HIDDEN

SPECIMEN FOR THE BOLTED JOINT


I

Fig. 16. Geometry of the hole and washer. Fig. 17. Finite-element model for the bolted joint.
Optimum bolted joints for hybrid composites 339

criterion, which is depicted in the following [30] 21


0
0
0
0 0

(?,‘+( y)‘+( Fr>l for a,,>0 18 -


0

15 -
Pa)
or

(T)‘+( Frtl for a,,<0 Pb)


6 -
where 2, is the interlaminar tensile strength,
and R and T are the interlaminar shear q First Peak Point
3 -
strengths in the l-3 and 2-3 directions, respec- o Maximum Load

tively. The material properties in Table 1 were


used and the interlaminar tensile strength was 0
I I I I
C G A
assumed to be equal to the transverse tensile E F D
Stacking Sequence Type
strength, Y,. The interlaminar shear strengths R
and T were assumed to be equal to the in-plane 2.1
(a>

shear strength, S, and the out-of plane shear


moduli, G13 and Gz3, were assumed to be equal
1.8
to the in-plane shear modulus, G,.
The load condition of stress analyses was the
1.5
cosine distribution, of which the x-directional
resultant force was 10 kN. 0
z
The hybrid composites of [OG/+ 45C/90G]s a 1.2 0 0
E 0
with seven different stacking sequences were
2
analyzed and compared with the experimental ’
r=
0.9
,
results. The analysis results predicted the first 2

peak load (A in Fig. 4), however, it could not 0.6


predict the maximum failure load. Figure 18a
and b show experimental and FEM results.
0.3
For a more accurate prediction of the joint
strength it might be necessary to include the
0
effects of the material non-linearity, the friction I I I I

E F D C G A
between the washer and the laminate, as well as Stacking Sequence Type
between the bolt and the laminate, and the stiff-
(b)
ness reduction due to failure during analysis.
Fig. 18. Comparison of experimental results with the FEM
calculated failure index result of the joints when the
external load was 10 kN ([OG/f45C/90G], (OG, 90G:
CONCLUSIONS 17.9%, f45C: 64.2%)). (a) Experimental results. (b)
FEM results.
In this work, the optimum conditions of bolted
joints for the hybrid composite materials com-
of the laminate under the washer. At the
posed of glass-epoxy and carbon-epoxy under
maximum failure load, delamination on the
tensile loading were investigated. The design
loaded side of the hole and the extensive
parameters investigated were ply angle, stacking
buckling of plies at the washer outside edge
sequence, the ratio of glass-epoxy to carbon-
were found.
epoxy, the outer diameter of the washer and the
2. From the static test results of the hybrid
clamping pressure. From the test results the fol-
composites with stacking sequences of [OC/
lowing conclusions were made.
+ 45G/_+ 45C/9OC], and [OG/ + 45C/90G],
1. A peak load occurred before the maximum (C: carbon-epoxy, G: glass-epoxy), it was
failure load, which was due to delamination found that the bearing strength increased as
340 J. H. Oh, k: G. Kim, D. G. Lee

the k45 plies were distributed evenly in the 6. Reinhart, T. J., Composites - Engineered Materials
F$m$o;k, Vol.1. ASM International, 1987, pp.
thickness direction, irrespective of the ratio
of glass-epoxy to carbon-epoxy and the 7. Vinson, J. R. & Sierakowski, R. L., The Behavior of
stacking pattern. Structure Composed of Composite Materials. Martinus
3. The bolted joint of [+45C/-45c/+45c/ Nijhoff, 1989, pp. 239-83.
8. Matthews, F. L., Joining Fibre-reinforced Plastics. Else-
(OC),/ - 45C/+45C/ - 45C/(90G),]s, which vier Applied Science, London, 1987.
has a 35.5% volume fraction of glass-epoxy, 9. Collings, T. A., On the bearing strength of CFRP
yielded the highest bearing strength. In this laminates. Composites, 1982, 13, 241-252.
10. Eriksson, I., On the bearing strength of bolted
case, the bearing strength increased as the
graphite/epoxy laminates. J. Composite Mater., 1990,
clamping pressure of the bolt increased up to 24,1246-1249.
71.1 MPa, then the bearing strength satu- 11. Quinn, W. J. and Matthews, F. L., The effect of stack-
rated to a constant value. When the washer ing sequence on the pin-bearing strength in glass fibre
reinforced plastics. J. Composite Mater, 1977, 11,
outer diameter was 20 mm, the failure mode 139-142.
was changed from bearing to tension. There- 12. Herrington, P. D. and Sabbaghian, M., Effect of
fore, the stacking sequence of [+45C/ radial clearance between bolt and washer on the bear-
-45C/+45C/OG,/-45C/+45C/-45C/90G,], ing strength of composite bolted joints. J. Composite
Mater, 1992, 26, 1826-1843.
would be the most suitable one. Also the 13. Godwin, E. W. and Matthews, F. L., A review of the
clamping pressure should be larger than strength of joints in fibre-reinforced plastics. Part I.
70 MPa and the outer diameter of the mechanically fastened joints. Composites, 1980, 11,
155-160.
washer should be 20 mm, respectively.
14. Collings, T. A., The strength of bolted joints in multi-
4. The finite-element analysis predicted the first directional CFRP laminates. Composites, 1977, 8,
peak load, however, it could not predict the 43-54.
maximum failure load. For a more accurate 15. Smith, P. A., Pascoe, K. J., Polak, C. and Stroud, D.
O., The behavior of single-lap bolted joints in CFRP
prediction of the joint strength it might be laminates. Composite Struct., 1986, 6, 41-55.
necessary during the analysis to include the 16. Kim, R. Y. and Whitney, J. M., Effect of temperature
effects of the material non-linearity, the fric- and moisture on the pin-bearing strength of compo-
tion between the washer and the laminate, site laminates. J. Composite Mater, 1976, 10,149-151.
17. Akay, M., Bearing strength of as-cured and hygro-
and between the bolt and the laminate, and thermally conditioned carbon fiber/epoxy composites
the stiffness reduction due to failure. under static and dynamic loading. Composites, 1992,
23, 101-108.
18. Waszczak, J. P. and Cruse, T. A., Failure mode and
strength prediction of anisotropic bolt bearing speci-
ACKNOWLEDGEMENTS mens. J. Composite Mater, 1971, 5, 421-425.
19. Agarwal, B. L., Static strength prediction of bolted
This research was supported financially by the joints in composite material. AZAA J., 1980, 18,
Korean Ministry of Science and Technology 1371-1375.
20. Soni, S. R., Failure analysis of composite laminates
under the international collaboration research with a fastener hole. In Joining of Composite Materials,
between KAIST and France (FRAMA- ASTM STP 749, ed. K. T. Kedward. American Society
SOFT+CSI). for Testing and Materials, 1981, pp. 145-64.
21. Chang, F. K., Scott, R. A. and Springer, G. S., Failure
of composite laminates containing pin-loaded holes -
method of solution. J. Composite Mater, 1984, 18,
REFERENCES 255-278.
22. Chen, W. H., Lee, S. S. and Yeh, J. T., Three-dimen-
1. Mallick, P. K., Fiber-reinforced Composites. Marcel sional contact stress analysis of a composite laminate
Dekker, New York, 1988, pp. 3-4. with bolted joint. Composite Struct., 1995, 30,
2. Schwartz, M. M., Composite Materials Handbook. 287-297.
McGraw-Hill, New York, 1984, Chap. 7. 23. Graham, U., Wisnom, M. R. and Webber, J. P. H., A
3. Reugg, C. & Habermeir, J., Composite propeller novel finite element investigation of the effects of
shafts design and optimization. In Advanced Compo- washer friction in composite plates with bolt-filled
site Material Proc. ICCM 3, Vol. 2, ed. A. Bunsell et holes. Composite Struct., 1994, 29, 329-339.
al., Pergamon Press, Oxford, 1980, pp. 1740-55. 24. Smith, P. A., Ashby, M. F. and Pascoe, K. J., Model-
4. Lee, D. G., Sin, H. C. and Suh, N. P., Manufacturing ling clamp-up effects in composite bolted joints. J.
of a graphite epoxy composite spindle for a machine Composite Mater, 1987, 21, 878-897.
tool. Ann. CZRP, 1985,27, (1) 365-369. 25. Hancox, N. L., Fibre Composite Hybrid Materials.
5. Lee, D. G., Kim, K. S. and Kwak, Y. K., Manufactur- Applied Science, London, 1981.
ing of a SCARA type direct-drive robot with graphite 26. Matthews, F. L., Roshan, A. A. and Phillips, L. N.,
fiber epoxy composite material. Robotica, 1991, 9, The bolt bearing strength of glass/carbon hybrid com-
219-229. posites. Composites, 1982, 13, 225-227.
Optimum bolted joints for hybrid composites 341

27. Kretsis, G. and Matthews, F. L., The strength of bol- 29. Edwards, K. S., Jr & Robert, B. McKee, Fur&amen-
ted joints in glass fiber/epoxy laminates. Composites, tals of Mechanical Component Design. McGraw-Hill,
1985,16,92-102. New York, 1991.
28. Bickford, J. H., An Introduction to the Design and 30. Chen, W. H. and Lee, S. S., Numerical and experi-
Behavior of Bolted Joints. Marcel Dekker, New York, mental failure analysis of composite laminates with
1981. bolted joints under bending loads. J. Composite Mater,
1995,29,15-36.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy