Fatigue

and
Residual Strength
of
Composite Aircraft Structures

Tonny Nyman
Department of Aeronautics
Kungliga Tekniska Hgskolan
(Royal Institute of Technology)
SE-10044 Stockholm
Sweden

Report No. 99-26

ISSN 0280-4646

PREFACE
Most of the work presented in this thesis was carried out within a national research program in
Sweden, Life Prediction Certification and in Service Assessment of Composite Structures,
between July 1995 to July 1998 at Saab AB and at the Royal Institute of Technology,
Department of Aeronautics. It was financially supported by Saab and NFFP, the National
Aeronautical Research Program in Sweden, for which I am very grateful.
Most of the work carried out in this dissertation stems from research ideas formulated within
that program in which I had the privilege to act as a project leader. Participating parts where
besides Department of Structural Strength and Loads Technology at Saab AB, the Aeronautical
Research Institute of Sweden (FFA) and the Royal Institute of Technology, Department of
Aeronautics (KTH).
I would like to thank my supervisor Professor Anders Blom at the Royal Institute of
Technology for his support and encouragement during the course of this study and Professor
Jan Bcklund for the opportunity to carry out this work at his department. Joakim Schn, at the
Aeronautical Research Institute of Sweden, is also greatly acknowledged as a friend,
companion in the project and as a valuable discussion partner. At the same institute I would
also like to thank Dr. Brje Andersson for his enthusiasm and deep knowledge, especially in
computational mechanics, which enabled us to perform state-of-the-art calculations with his
developed hp-adaptive FE-system STRIPE. I am also very grateful for the support received
from colleagues at Saab AB, especially Tomas Ireman, Christina Altkvist, Magnus Friberg and
Anders Bredberg at the composite group but also to Hans Ansell, previous manager of the
department and co-author of two of the papers. To my friends in the metallic group, although
you havent contributed to the appended papers I have benefited from your friendship for which
I am very grateful. A special thanks to a skilful, experienced and observant technician at CSMmaterialteknik, Bernt Forslind.
Finally I would like to thank my family, Ingrid, Torbjrn, Per and Henrik for their support and
patience during this period. To my very special dog, XITA, thank you for forcing me out into
long walks in the forest, thus clearing my thoughts and giving me new fresh angles to current
problems.

Linkping, August 1999

T. Nyman

ABSTRACT
This thesis consists of six papers within the area of composite fatigue and residual strength in
aircraft structures. The dissertation is a mixture of experimental, numerical and theoretical
work with focus on some structural hot-spots as defined in the certification process.
Basic fracture mechanical properties were investigated, both numerically and experimentally
for the standard DCB-and ENF-specimens, with respect to different ply interfaces and materials. Toughness measurements on specimens with non-zero interfaces have a considerable
higher toughness than specimens with zero interfaces, due to the fact that the initial crack deviates from the original symmetrical crack plane. A good prediction between FE-calculations and
beam data analysis methods could be obtained if the non-homogenous characteristics of specimens with non-zero interfaces was taken into account. The cyclic delamination growth rates
increased rapidly with increased G, raising doubts about the structural applicability of such
growth laws. The problem is rather a threshold problem. The importance of repeated loading
on composite structures with artificial delaminations is also demonstrated.
The rule of interlaminar stresses with respect to damage initiation on notched and unnotched
specimens were numerically investigated with an state-of-the-art hp-adaptive FE-technique and
compared to some existing semi-analytical models. Though the number of initiation points for
various quasiisotropic layups were found to be quite different the initiation strain, from a calculation point of view, was almost the same. Investigated semi-analytical models as formulated
today were not able to cope with the complex stress-state occurring at free edges.
A simplified approach for composite fatigue design is suggested. The methodology is based on
fatigue failure functions established on a laminate level and thus the calculated uniaxial properties will reflect the overall behaviour of the laminate. To establish the fatigue failure functions, a given SN-curve for a multi-directional laminate is considered as a series of failure
points where the failure in the short life-region essentially is assumed to be fibre fracture and
in the long life-region is assumed to be controlled by matrix properties.
An experimental investigation was also carried out regarding effects on resulting life for composite structures. Different elimination levels were considered on various load spectra and on
constant amplitude block loading. Results revealed that the elimination level could be set to
approximately 50% of the maximum load range. This means that considerable time and cost in
structural verification testing can be obtained. Spectrum fatigue testing accounting for a characteristic number of cycles could be mapped on constant amplitude data which means that a life
prediction methodology can be radically simplified.
The so called soft inclusion theory and delamination buckling theory was investigated on
impact damaged composite structures for residual strength prediction. The investigation was
supported by FE-technique for determination of the stress distribution in the buckled state and
for characterisation of the damaged region. It could be concluded that for low-energy impact,
although with conservative assumptions on stiffness reduction, the soft inclusion theory is unconservative for residual strength prediction. In contrast the delamination buckling theory
shows good agreement for various impact energy levels, thicknesses and layups.

ii

CONTENTS
PREFACE ............................................................................................................... i
ABSTRACT ........................................................................................................... ii
CONTENTS ......................................................................................................... iii
DISSERTATION .................................................................................................. iv
INTRODUCTION ................................................................................................. 1
1

Background .................................................................................................. 1

1.1

Certification process ................................................................................ 2

1.2

Damage characterisation .......................................................................... 5

Interlaminar stresses ..................................................................................... 7

Basic fracture mechanical properties ............................................................ 9

Composite fatigue ....................................................................................... 14

4.1

Basic methodology ................................................................................ 14

4.2

Truncation and elimination .................................................................... 16

Residual strength ........................................................................................ 16

5.1

Equivalent damage methodology .......................................................... 18

5.2

Delamination buckling theory ............................................................... 20

5.3

Artificial delaminations ......................................................................... 21

SUMMARY OF APPENDED PAPERS ............................................................. 22

DIVISION OF WORK BETWEEN AUTHORS ................................................ 25
RECOMMENDATIONS FOR FUTURE WORK .............................................. 26
REFERENCES .................................................................................................... 27
PAPER A .....................................................................................................A1-A29
PAPER B ...................................................................................................... B1-B35
PAPER C ...................................................................................................... C1-C21
PAPER D ......................................................................................................D1-D23
PAPER E........................................................................................................E1-E24
PAPER F........................................................................................................ F1-F22

iii

DISSERTATION
This dissertation contains the following appended papers:

PAPER A:

Schn J., Nyman T., Blom A. and Ansell H., A numerical and experimental investigation of basic fracture mechanical properties in the DCB-specimen. Accepted for publication in Composites Science and Technology.

PAPER B:

Schn J., Nyman T., Blom A. and Ansell H, A numerical and experimental investigation of basic fracture mechanical properties in the ENF-specimen. Submitted to Engineering Fracture Mechanics.

PAPER C:

Nyman T. and Friberg M., Interlaminar stresses in composite notched and

unnotched laminates. Accepted for publication in Journal of Reinforced
Plastics and Composites. Presented at the 13th International Conference of
the American Society for Composites in Baltimore, September 21-23,
1998.

PAPER D:

Nyman T., " Composite Fatigue Design Methodology A Simplified

Approach". Published in Composite Structures 35 (1996) 183-194. Condensed version of paper presented at the Third International Conference on
Composites Engineering, ICCE/3, in New Orleans, July 21-26, 1996.

PAPER E:

Nyman T., Ansell H. and Blom A., Effects of truncation and elimination
on composite fatigue life. Submitted to Composite Structures.

PAPER F:

Nyman T., Bredberg A. and Schn J., Equivalent damage and residual
strength for impact damaged composite structures. Accepted for publication in Journal of Reinforced Plastics and Composites. Presented at the
13th International Conference of the American Society for Composites in
Baltimore, September 21-23, 1998.

iv

INTRODUCTION
1

Background

New aircraft must comply with requirements for durability and damage tolerance. A modern
combat aircraft today contains high percentages of advanced composite material in flight critical primary structures, e.g the B-2 stealh bomber, the F-22 advanced tactical fighter, the F-117A
stealth fighter and the V-22 tilt-rotor, Ref. [1]. The new Swedish fighter JAS39 Gripen utilises
about 20%, by weight, carbon fibre/epoxy composite laminates in the primary structure, see
Fig. 1.
Carbon-Fiber Composite (CFRP)

Fin tip

Glass-Fiber Composite (GFRP)

Vertical fin

Aramid-Fiber Composite (AFRP)

Fillet

Equipment door

Main wing

Radome

Canard

Leading edge flaps

Nose landing
gear doors

Access
Doors
Main landing
gear door

Figure 1 The extent of composite materials in JAS39 Gripen.

Specific damage tolerance requirements and design criteria for advanced composite
materials were not specified during design and development of various military aircraft introduced in the 1970s and early 80s such as the F-16, F-15 and F/A-18, Ref. [1]. Instead, damage
tolerance were demonstrated by extensive structural testing and by the use of low design strain
allowables in the design. An excellent review of lessons learned with the use of advanced composites can be found in Ref. [2]. Some conclusions are cited below:
o Significant weight savings have been achieved
o Outstanding fatigue and corrosion resistance have been demonstrated
o In-service performance has been very successful
o Durability of thin skin honeycomb structures has been less successful
A review of composite design philosophy/process can be found in Ref. [3] where it is empha1

sized that a good design is usually one where complexity and sophistication are kept to a minimum, as these often only serve to increase costs and reduce reliability. It is also pointed out
that joints frequently have been a problem, thus design which will reduce the number of joints
will generally lead to weight and cost savings and an improvement in component integrity. Considerations when looking at detailed design can be summarised as follows:
o Avoid offset loads which result in local bending
o Avoid peeling forces
o Consider galvanic corrosion effects between carbon composite and many metals
o Look for local weaknesses and anticipated failure modes
o Remember that joints may be locally stiffened by the addition of extra laminations
Of special concern are also what type of loads that are significant/characteristic and how they
are applied on the geometrical and laminate configuration at hand, i.e whether they are tensile,
compressive, shear etc., and how they appear in the sequence.

1.1

Certification process

Currently, the main design drivers for most composite structural components is either stiffness,
static strength or low energy impact damage. In order to avoid structural problems current
design procedures have led to rather low allowable strain levels. It is anticipated that these values will be increased in future aircraft and then it may be necessary to further consider the likelihood for fatigue damage. In the certification process, durability and damage tolerance criteria
require composite structures containing undetected damage to be acceptable to fly.
Current design philosophy is based on no damage growth criteria and is realised by conservative strain limitations (e.g assuming that a 6 mm hole can be present everywhere in the
structure representing e.g impact damage, manufacturing defects etc.) and verification testing
on components and built-up structures. Consequently, expensive testing, not only for certification of new structures but also for comparable structures with only minor alternations, is frequently performed by the industry. The problem is usually represented by a test pyramid where
each stage refers to an investigation level in terms of specimen category, see Fig. 2.

MATERIAL DATA BASE

Generic Specimens

FULL SCALE
COMPONENTS

SUBCOMPONENTS

STRUCTURAL FEATURES
Non - Generic Specimens

DETAILS

ELEMENTS/
LAMINATES

COUPONS

Figure 2 Test pyramid to certify composite structures.

In Fig. 3 some of the structural verification tests performed on JAS39 are shown. In fact the testing is on-going and extended with 39B the two-seater.

MIDPOINT HINGE
INBOARD
ELEVON

FATIGUE

CANARD
PIVOT
ATTACHMENT

RUDDER

WINGFUSELAGE
JOINT

LEADING EDGE
FLAP INBOARD
GEAR BOX

86

INBOARD
HINGE OUTBOARD
ELEVON

AFT FUSELAGE
FIN & RUDDER

87

COMPLETE
AIRFRAME

AFT ROOT RIB

OUTBOARD
FLAP

85

CANOPY &
ATTACHMENT

88

90

89
GUN FORWARD
ATTACHMENT

FATIGUE & DAMAGE TOLERANCE

DEFLECTOR
FORWARD
ATTACHMENT

91
MIDPOINT HINGE
INBOARD
ELEVON
UPPER AFT
WING-FUSELAGE
FITTING
WINGFUSELAGE
JOINT

92

93

CANARD
& PIVOT

OUTBOARD
ELEVON

INBOARD
ELEVON
CANARD
PIVOT
ATTACHMENT

ENGINE AFT
MOUNT
AFT ROOT RIB

DAMAGE TOLERANCE
INBOARD
HINGE OUTBOARD
ELEVON

AFT FUSELAGE
FIN & RUDDER

Figure 3 Structural verification testing on JAS39 Gripen.

Special attention must be made regarding coupon v.s structural testing. For example,
edge delamination occurs in unnotched coupons and significantly influences the result. However, for a structural panel the effect is much less. Also of special interest is, depending on e.g
geometry, the difference in static and cyclic strength between tension and compression loading.
3

Of special importance is also the scatter in structural response achieved in structural tests, i.e
the structural response variability. Besides the scatter on a coupon level several other factors
contribute to the overall scatter, e.g structural geometry, loading conditions, manufacturing and
material conditions. Due to this scatter unexpected failure can occur, i.e hot-spot failure, see
Fig. 4. Very often these hot-spot failures are associated to out-of plane loading emphasizing
the need for analysis capability with respect to interlaminar stresses.

Strength Distribution
F(P1)

Mean Strength

Hot-Spot Failure
Design allowable
Structural Response
Distribution
p()

Mean Structural Response

Observed Structural Response
P1
Applied Load

Figure 4 Schematic of Hot-Spot failure for a static test, from Ref. [4, 5].
In the process of structural verification testing usually both composite and metallic parts
are involved. In order to have reasonable test times, truncation and elimination of the load
sequence must be done. This process, load sequence reduction, usually leads to different conclusions for the composite structure and for the metallic part.
Lessons learned within the certification process are, see e.g. Ref. [2]:
o There is a need for analysis capability within the area of interlaminar stress calculation since secondary loads (out-of-plane loading) are almost impossible to eliminate
from complex built-up structures.
o The full scale static test is the key test for composites since it is the primary hot spot
indicator.
o Lack of reliable life prediction tools makes it necessary to rely on a no-growth concept and makes it impossible to investigate the consequences of a load spectrum
update.
o There is an urgent need for an agreement what should be a detectable size of an inden-

tation, i.e there is a lack of appropriate definition of barely visible impact damage
(BVID). There is also need for an agreement of cut-off energy, values from a few
Joules to 140 J exists.
o There is also need for rational repair/reject criteria including a methodology for
determining inspection intervals.
o No complete airframe test validation is required for composite damage tolerance certification. Extensive testing has shown that subcomponent validation tests accurately
represent fullscale composite damage tolerance behaviour.

1.2

Damage characterisation

During operating conditions damage of different nature, e.g service mechanical damage,
fatigue damage or environmental damage, will be observed/detected. Facing different damages,
manufacturing or service induced, delamination related damage is one of the far most important
failure mechanisms of aircraft-composite structures and can be detrimental for flight safety.
Consequently delaminations must be considered in the design process but also in structural verification testing. Examples are delamination growth due to local buckling, delamination growth
from free edges and notches such as holes, delamination growth from ply drops, and impact
damaged composites containing considerable delamination. Delamination growth can also
occur due to interlaminar stresses which can arise from fuel pressure variations and in complex
structures due to unanticipated loading.
The complex nature of composite failure, involving different failure modes and their
interactions, makes it necessary to characterise / identify the relevant parameter(s) for damage
resistance, accumulation and life prediction. It is not obvious that a major characteristic failure
mechanism on a coupon level will be equally interesting for a real structure. It is well recognised that the most damaging state is impact damage at least for compression dominated loading. In the certification process different damage states must be identified to enable verification
of the structure. For instance for barely visible impact damage (BVID), the structure must sustain ultimate load before and after cyclic loading.
The three main types of defects observed in fiber composites are transverse cracking,
splitting, delamination between plies and fiber fractures which occur mainly in the last stage of
the life of the structure Refs. [6], [7]. Transverse matrix cracks are initiated very early in the
fatigue process and they multiply until a saturation state is reached, Characteristic Damage
State (CDS) Ref. [8]. The loading level has a large effect on the rate of crack multiplication Ref.
[6], [9].
The matrix can not easily deform in a non-linear way which makes it difficult to relax any
stress concentrations. Instead, matrix cracking will serve to reduce any stress concentrations
and the consequence will be redistribution of load in the composite and increased stress and
strain in the 0o plies. Most of the matrix cracks propagate through the thickness of the ply but
they can usually not propagate into the next ply if the difference in fiber orientation is sufficiently large. A stress singularity will occur at the crack tip of the matrix cracks which during
5

fatigue loading might induce delamination at the interface. At those points the stress field can
be expected to be very complex.
In the 0o plies matrix cracks will form parallel to the fibres, i.e. splitting, and the distance
between the splitting cracks is larger than for the transverse matrix cracks and it is randomly
distributed throughout the specimen, Refs. [7], [9]. These dispersed cracks are observed mainly
at low loads and increase with the number of cycles up to failure. At higher stresses, localized
longitudinal cracking appears to be the main damage mechanism prior to failure, Ref. [9]. The
composite component will fail when the 0o plies fail, requiring fiber fracture. It has been
observed that most of the fiber failure and damage localization occurs in the last 10% of the life
of the laminate, Ref. [9]. The last part of a specimens fatigue life has been reported to show an
accelerating rate of damage development characterized by the localization of damage in zones
of increasing crack interaction, Ref. [10]. This involves the coalescence of microcracks and
development of macro-cracks.
It has been found that the growth of a delamination is dependent on the stacking sequence.
Delaminations were found to propagate perpendicular to the loading direction in a (0/903)s laminate, as opposed to parallel to the loading direction in a (0/902)s laminate. The delamination
area can become very large at some interfaces, values of 57% delamination have been reported.
This means that towards the end of the fatigue life of the composite the different plies are more
or less isolated from each other. It is also possible to have delaminations to form groups with a
high density of delaminations Ref. [7].
Compression-induced damage is of a different nature than tension-induced damage. The
compression induced damage will depend on the interlaminar stresses which in turn strongly
depends on the stacking sequence and the thickness of the laminate. Tests were conducted by
Ratwani, Ref. [11], on AS/3501-6 graphite epoxy with four different stacking sequences. All
laminates had the same percentage of 0o, 90o and 45o degree plies. Constant amplitude tests
were performed at a loading ratio of R = 1, . Depending on stacking sequence different
failure modes were obtained for the central circular notched specimens. It was also concluded
that the damage growth changed direction and location completely when the stacking sequence
was changed. The location of delamination and matrix cracking for the different laminates coincide with the highest interlaminar shear or normal stress calculated with FE-technique.
Test results presented, especially compressive test results, must be critically reviewed
since interlaminar stresses, depending on failure mode, can have played a dominant role, Ref.
[12]. There is always a question to what extent they are invoked in the test results, i.e. they must
be viewed in light of test methods, for example special fixtures used to prevent buckling, see
e.g. Highsmith, Ref. [13]. Compression-compression fatigue damage in notched ((45/02)3)s
composite laminates, T800/924C, consists of matrix cracking, delamination, and at applied
compression stresses approaching the laminate strength, fibre microbuckling, Ref. [14]. Axial
loading tests were performed at constant-amplitude, R=10, with a frequency of 10Hz.
A failure prediction technique which takes microbuckling into account was introduced by
Soutis et al., Ref. [15]. Failure is due to 0o fibre microbuckling surrounded by delamination. The
microbuckling zone initiates at the hole boundary. It extends stably under increasing load
6

before becoming unstable at a critical microbuckling length of approximately 2-3 mm. A

microbuckling failure criterion, in conjunction with Lekhnitskiis complex variable technique
Ref. [16], was introduced by Grdal et al., Ref. [17]. The schematic kinking mode of failure is
shown in Fig. 5.

Figure 5 Kinking mode of failure, Ref. [17].

In-service detected damage must be correlated to life prediction methodology and inservice inspection tools and repair/reject criteria must be developed. There must also be a correlation between what is inspectable/seen and life prediction- residual strength methodology. A
mutual correspondence between methods and characterized damage must be sought.

Interlaminar stresses

Secondary loads are almost impossible to avoid in a built-up structure. These secondary loads
will produce out-of-plane loading resulting in interlaminar stresses in principal structural elements which could cause premature failure. It is believed that these interlaminar stresses play
a dominant role in the process of delamination and thus are a limiting factor for the life of a
structural component. Interlaminar stresses will arise when e.g a structural component is subjected to acoustic random loading, post buckling, fuel pressure variations or when impacted.
These stresses must thus be considered in e.g. structural verification testing and in the design
process.
Interlaminar stresses also provide important information regarding damage initiation/
growth and could therefore be the key for better understanding of the failure process. The damage initiation will most likely depend on these interlaminar stresses and consequently on the
stacking sequence. It is also likely that delaminations start to develop where interlaminar
stresses are most severe.
To have analysis capability within this area, i.e to calculate the resulting interlaminar
stresses in a straightforward manner is important. Some out-of-plane sources are exemplified
in Fig. 6 below. Coupled with a proper failure criterion the calculations could also be an impor-

tant design tool.

Buckling

Hole

Free edge

Beam Radius
Bond Joint

Ply drop-off
Local bending

Figure 6 Sources of out-of-plane loads.

Among the first persons to identify and recognise the existence of interlaminar stresses
was Pagano, Ref. [18]. The three-dimensional problem was solved by applying finite difference
techniques to the governing elasticity equations. The existence of a boundary layer and the possibility of a singular behaviour was launched. Since then many three-dimensional finite element
solutions have been published see e.g. Refs. [19, 20] and [21]. The singularity problem is e.g
discussed in Refs. [22] and [23]. Boundary collaction methods, methods depending on damage
theory, specially developed continuous strain high order elements but also plasticity related
methods have been used for the solution of the problem, see Refs. [24, 25, 26, 27]. Though the
FE-method provides relative accuracy to the problem, the need for simplified methods is obvious. In 1986 an article which fulfilled that need, Ref. [28] was published. By assuming stress
functions and applying traction continuity and minimising the complementary potential function the problem could be solved. The article was followed by a number of articles dealing with
other types of loading not only uni-axial loading but also combination of thermal loads and
mechanical effects, Refs. [29, 30, 31, 32]. Basically the same theory was adopted to study the
influence of a hole on the interlaminar stresses, i.e. a curved free edge Refs. [33, 34, 35, 36, 37].
By turning the problem into a series of straight free edges an approximate solution could be
found. FE-based solutions to the hole problem can be found in e.g. Ref. [38].
For the prediction of damage initiation stress based criteria, see e.g. Ref. [39, 40], or fracture mechanics related criteria, see Ref. [41], have been adopted. Delamination and failure analysis are also discussed in Refs. [42, 43, 44, 45]. An excellent review has been made in Ref. [46]
which also contains a number of numerical solutions to the free edge problem.

Normal and or shear interlaminar stresses arise at the free edges depending on e.g. the
mismatch of properties between the layers. The magnitude of the stresses will depend on stacking sequence and on external loading. Interesting is also that these stresses will depend on temperature, as layers have different expansion coefficients, and also with moisture content, as
layers expand to different extents on absorbing external moisture, Ref. [47].
In paper C interlaminar stresses for notched and unnotched plates, for various stacking
sequences, are calculated with an hp-adaptive FE-analysis technique and compared to some
existing semi-analytical methods. A delamination criterion is also examined and compared to
calculated results. Fig. 7 shows, for one of the analysed cases, the interlaminar stresses at the
hole-edge.

rz

rr

zz
z
r

1%

z at the hole at theta=67.5

Taothetaz
r=3 layup [90/45/0/45]

3s

500

500

Size of fiber diameter

1000

1500

2000

Coarse P=5
Refined P=8
0

0.13

-450

0.26

00 450

0.39

0.52

0.65
0.78
0.91
Thickness, z [mm]

900 -450

00

450

1.04

1.17

900 -450

00

1.3

1.43

1.56

450 900

Figure 7 Interlaminar stresses at an cut-out.

Basic fracture mechanical properties

It is important to improve the knowledge of delamination growth both theoretically and experimentally. The most straightforward way is by studying simple test specimens, such as the dou9

ble-cantilever beam (DCB) or the end notch flexure specimen (ENF). It is generally believed
that the mode separated energy release rate of delamination growth is a good measure of the
composite materials resistance to delamination growth. Delamination growth can be separated
into two different types, quasi-static loading, e.g due to very large accidental overloading, and
fatigue loading. By using the DCB-specimen or the ENF-specimen the mode I and mode II
energy release rates can be measured and this information can then be used for comparing different composite materials and to increase the understanding of delamination growth in composite materials.
Several round robin tests of unidirectional composites have been performed, with focus
on e.g. delamination onset, T-tab v.s hinge, insert type and thickness, Ref. [48]. These investigations have been used for establishing the ASTM (American Society for Testing and Materials) DCB standard. An overview of experiments and analyses being performed at the coupon
level within the GARTEUR (Group of Aeronautical Research in Europe) action group AG 16
can be found in Ref. [49]. The crack initiation has been studied with microfocus radiography
on DCB-specimens and it was found that the crack initiates in the interior of the specimen
resulting in a curved crack tip, and that the crack initiation corresponds to the non-linear loaddisplacement curve, Ref. [50]. For graphite/epoxy composite specimens the interfacial ply orientation do not have a significant effect on GIc values, Refs. [51], [61].
The simple beam theory has difficulties simulating the deformation near the crack tip.
More advanced models have been introduced by Olsson, Ref. [52], and Sun-Pandey, Ref. [53].
In a comparison with a FEM solution the FEM results lie between the results obtained from the
Olsson and Sun-Pandey models, see Fig. 8.

Q
1Q
5m
2146-_57D
o8O
)(/6_1D
cm
)sn(o_c|m
etvnfsS
_|let/s1S
_T
hem
/ty10soT
hkm
w
y0oicrkw
0R
neilfcrot-0R
lesca-1leso92m
nodm
nofidtnfitanan 450/450
Olsson

Q
m
15Q
6-)B
7_25D
om
(41c)0_5Q
o6-4D
e/|anvm
_2ftc5so(4S
_/)|0etm
1o(T
sD
_hnvtyS
fe0/cosm
1kT
_w
i0h|cetryosnR
fm
k0ltS
_w
ie/o-cr1asT
ltR
m
0hcyosne-2fk0lm
tw
i9creoasdlR
10ncm
fios-2tedfn9oiatnm
fodafnitfan 450/450
Beam

0
0
Q
1Q
5m
2196-_07D
o8O
)5(/6_1D
cm
)sn(o_c|m
etvnfsS
_|let/s1S
_T
hem
/ty10soT
hkm
w
y0oicrkw
0R
neilfcrot-0R
lesca-1leso92m
nodm
nofidtnfitanan 90 /90
Olsson

0
0
Beam
B
Q
154m
ea6192-_07D
o)5(/_c0m
(o_|entvfsS
_et/1T
hse_m
ty0okw
0oicr0R
nelfot-ecsal1o92m
nodnfitfan 90 /90

Q
594m
162-m
_D
o0164)5O
(c-9m
_0o/s5n|eo(fvt_S
nos/lt1efvhT
_m
y0stoekw
0oicrne0fR
lotc-aelsnft1s9coal2m
nodnfitfan 00/00
Olsson

0 0
Beam
Q
Q
5m
196-2_B
)Q
D
5o3(_8)cQ
m
D
041o_)|c6m
e-nav2tf4/sD
_S
o|0e9t(4)s1cD
m
S
_T
ho/y10|cm
_T
envhtkfosyw
0S
|ic_er/tkos1R
w
n0S
T
ilfhc_rm
/yt1-R
oe0T
lshakm
cyw
1-ioersk2R
w
n09ilm
fcerot-R
d0lm
sancfi1o-dest2nfi9m
toandm
fiaondtfitanan 0 /0

1,08
1,07

G(Beam models)/G(Stripe)

1,06
1,05
1,04
1,03
1,02

200

150

100

1,00

50

1,01

Cracklength (mm)

Figure 8 Comparison between Olssons model and classical beam theory.

In order to obtain consistent results when analysing experimental data it is necessary to
10

consider end rotation and deflection of the crack tip, the effective shortening of the beam due
to large displacements of the arms, and the stiffening of the beam due to the presence of the
blocks bonded to the specimens, Ref. [54].
The energy release rate along a straight crack will not be constant, but approaches zero or
infinity, depending on interface ply orientation, at the vertices where the crack intersects the
free surface. This will cause the crack to become curved and the angle the crack makes with the
2
surface is determined by the interface ply orientation, Ref. [69]. A ratio D C = D 12 ( D 11 D 22 )
has been introduced, Refs. [55], [56], and found to correlate with the difference in energy
release rate at a specimens centre versus that at its edges, Ref. [57]. A ratio B t = D 16 D 11 has
also been introduced and found to correlate with the asymmetry in the energy release rate distribution about the centre of the specimens width, Refs. [55], [56]. Since the energy release rate
is not evenly distributed along a straight crack the crack tip will become curved. This has been
studied analytically, Ref. [55], [57] and with FEM, Ref. [58].
For an ENF-specimen it has been observed numerically that the energy release rate distribution is uneven with a straight delamination front, with maxima occurring at the points
where the delamination front intersects the free surface. To predict this behaviour a non-dimensional ratio Dc has been defined where conventional laminate plate theory notation is used, Ref.
[59]. The amount of mode III energy release rate for ENF specimens increases with increasing
Dc ratio, Ref. [72]. The vertex singularity at the points where the crack intersects with the free
surface causes the energy release rate to approach infinity for a straight crack front. This will
cause subcritical crack advance at the specimen edges which qualitatively correlates with the
Dc ratio, Ref. [60]. During toughness testing of specimens with different interfaces only minor
influence on interface type has been found, Ref. [61].
An analytical equation based on beam theory taking friction into account can be found in
Refs. [62, 63]. Mall et al., Ref. [64], have studied the influence on energy release rate from friction, using a two-dimensional finite element model which assumed plain strain condition. The
energy release rate was computed with a virtual crack closure technique and it was found that
the friction effect is large for short crack lengths and negligible for longer crack lengths. Russell, Ref. [65], estimated the effect of friction from the extent of hysteresis observed during
loading and unloading of ENF specimens and friction was found to result in an overestimation
of the energy release rate of less than 2%.
The effect of fiber orientation at the interface for glass/epoxy have been studied by Hwu
et al., Ref. [66], and no clear trend was found. It has been found that the critical energy release
rate increases with a thicker matrix layer at the interface. Unidirectional specimens containing
a resin rich interface of 50 m and 200 m showed an increase in delamination resistance of up
to 200% when compared to control specimens of the same material. The increase in toughness
observed was attributed to the relief of plastic constraint at the crack tip, Ref. [67]. For the AS4/
3502 composite a damage zone, characterized by extensive microcracking, has been found in
front of the delamination. The damage zone was at least 180 m long and extended about 5 m
above and below the plane of the delamination, Ref. [68].
The distribution of the total energy release rate along the crack tip for various layups, with
11

the interface crack at the 0o/0o, 90o/90o, and 45o/45o interfaces are shown in Fig. 9. The crack
length is 25 mm and the coefficient of friction is assumed to be zero. The energy release rate
has been normalized with the total energy release rate (GII+GIII) at the centre of the specimens.
In the figure the energy release rate for mode II and III have been added. This can be motivated
by the similarities in fracture which might cause G IIC = G IIIC , Ref. [66]. The specimen with
a 90o/90o interface has the flattest distribution followed by the specimen with a 0o/0o interface
whereas the specimen with the 45o/45o interface has the most uneven distribution. The energy
release rate approaches infinity as the free surfaces are approached. This is due to the vertex
singularity located at those positions and has been described elsewhere, Ref. [69]. The uneven
distribution of the energy release rate would suggest that subcritical crack growth, starting from
the free edges, where the energy release rate approaches infinite values, will occur prior to
unstable crack growth. This has also been observed experimentally, Refs. [70] and [71].

Figure 9 Distribution of G along the crack front for an ENF-specimen, different layups
considered.
12

The classical data analysis methods are strictly valid only for homogenous beams. For
multidirectional laminates the introduced error will depend on the stacking sequence and this
has been discussed in detail in Ref. [72]. A more general expression for the compliance is introduced with an alternative expression for GII, i.e. a modified beam theory, where the ratio, R, of
flexural rigidities of the cracked and uncracked region is involved. For a homogenous isotropic
material R=8 but will alter for a multi-directional laminate depending on stacking sequence. As
can be seen in Fig. 10 the modified direct beam theory, according to that equation give reasonable predictions as compared to the FE-solutions. For a/L=0.5 the error is less than 5%. Increasing the crack length to 35mm (a/L=0.7) good predictions are made except for the 45o/45o
interface where approximately the same error remains, i.e 5%.

Figure 10 Comparison FE-model with modified direct beam theory

In paper A and B results from thorough investigations, both numerical and experimental, on the
DCB- and ENF-specimens are discussed. The numerical investigation has been done with an
hp-adaptive FE-analysis technique.

13

4
4.1

Composite fatigue
Basic methodology

There is a strong need to be able to predict the fatigue life for a structural component under realistic loading conditions in aircraft design. These conditions imply general in-plane loading
where the load sequence produced will contain numerous loading conditions, R-values, that
must be handled in a convenient way in a design methodology, see Fig. 11.

CRUISE
RETURN

CLIMB

AIR COMBAT

INTERCEPTION
DESCENT

CRUISE
Load factor - nz

LANDING

No of exceedences

Figure 11 Schematic of loading conditions.

Structural-verification testing with similar components but with different loading conditions
and repair/reject criteria are areas where an improved methodology for life prediction would be
beneficial. Composite fatigue, i.e life prediction, is not a straightforward matter depending on
various failure modes and their interactions. Scanning the literature, it is obvious that there is
no complete, ready to use, model today. What exists are different approaches to solve specific
problems, regarding e.g different R-values, layups, notched / unnotched behaviour etc.
One of the problems encountered with composite materials is that every new lay-up constitutes a new material, thus making the need for testing infinite. Besides from constraint
effects, causing differences between unidirectional and multi-directional laminates, there is
also a scale-up problem, i.e. going from coupons to a structural part will not necessarily yield
the same test results with respect to strength and fatigue life. The current philosophy is based
14

on no damage growth and this is achieved by conservative strain limitations and verification
testing on components and built-up structures, i.e. basically a threshold formulation. Any extension of that philosophy must also take into account current static design methodology and current NDT-techniques.
Pioneering work in the fatigue area of composite materials has been done by Schtz et.
al., Ref. [73]. A general overview directed towards aircraft structures can be found in Ref. [74].
For classic fatigue and especially on life prediction methodology an overview was made in Ref.
[75]. Life prediction of CFRP (Carbon Fibre Reinforced Plastics) has been investigated in the
past from different points of view. This would typically imply, degradation models, Refs. [7680], concepts with damage modelling including residual strength concepts, Refs. [78] and [8183] or semi-empirical models, Refs. [73, 74, 84-91]. Also typical is that many models are based
on micro-mechanical phenomena observed under unidirectional loading for unidirectional laminates. There is, however, little hope that damage/fatigue models built on laminae constituents
with associated micro-mechanical phenomena, e.g. micro-cracking, but not considering the
multi-directional nature of real structures would solve the problem of structural life prediction.
The nature of fracture is essentially ruled by, at least for in-plane loaded structures with loadcarrying fibres in all major directions, when essential fibre fracture starts to occur. A general
overview of the problem can be found in Refs. [11, 13, 47, 74, 98-104]. In Refs. [105-107] suggested methodologies for constant-amplitude, block and variable loading can be found.
In paper D a methodology is suggested which in a simplified manner takes into account
the laminate behaviour. That is the lamina characteristics are deduced from the global behaviour, see Fig. 12.

x
x

Figure 12 Schematic determination of fatigue failure functions, constant (average)

stress/strain.

15

4.2

Truncation and elimination

Durability and damage tolerance criteria are and become more and more important. From a
certification point of view there is a strict demand to validate the long-term behaviour of
composite structures during repeated loading. It is also absolutely necessary to know which
states of loads in the sequence that can be truncated or eliminated in structural verification
testing and under operational conditions. That is, which load states dominate/characterise
the life of a structural component.
An extensive investigation was conducted by Badaliance and Hill, Ref. [108, 109], on
effects of fighter attack spectrum on composite fatigue life. Typical wing spectra were investigated, where e.g clipping to 90% test limit stress, addition of stress overloads, addition of low
loads, elimination to 70% test limit load and clipping of tension loads, were investigated. Simple coupons with two different layups, zero- and matrix dominated were used. The importance
of retaining high loads in the spectra is emphasized, and cycles that have peaks less than 50%
of limit load can be eliminated since they do not contribute to the degradation. A comprehensive
investigation has also been made by Shtz and Gerharz, Ref. [73], with basically the same
research approach, i.e effect of truncation, overload and elimination on life. A typical zerodominated layup was used with un-notched specimens. The conclusion was that elimination of
low loads, roughly 70% was eliminated, was the prime effect of investigated parameters. It is
also well known that testing with basically un-notched specimens produce larger scatter than
specimens with e.g an open hole, Ref. [110]. In the certification process for the Navy F/A-18
E/F upgraded model MacDonnel Douglas shortened their composite spectra with the purpose
to reduce time and cost. Up to 60% of Test Limit Load (TLL) were eliminated, Ref. [111]. The
effect of elimination on a compression dominated spectrum, TWIST, was studied on notched
quasi-isotropic specimens by Phillips, Ref. [112], The results suggest that high loads should be
retained and that it is likely to have large reductions in the low-load end when defining test spectra.
A limited experimental investigation has been carried out and is discussed in paper E. It
was demonstrated that it is possible to map calculated and tested results from spectrum testing
on constant-amplitude data if a characteristic number of cycles were taken into account. The
experimental results also indicated that the elimination level can be set to approximately 50%
of maximum range for in-plane loaded structures.

Residual strength

Structural durability, damage resistance and damage tolerance are important tasks, not only in
the design process but also when the aircraft is in operational use. There are many sources and
types of damage e.g: fatigue cracking, environmental degradation, or damage introduced by
foreign objects. When occurring, all types of damage need immediate attention for determination of the effect on aircraft performance or functionality. There is therefore a need for simplified predictive methods for rapid assessment of occurring damage.

16

We are still some way from being able to predict the true effects of service damage. But
as pointed out in Ref. [113], if damage could be reduced to an idealised notch and/or delamination then it might be possible to relate these to local stress fields in a realistic structure and
therefore be able to predict structural failure loads. A schematic view of the actual problem can
be seen in Fig. 13.

Damage Resistance
Damage size
Structural
configuration

Operational
conditions

Interlaminar
stress calculation

Delamination
buckling

Hole damage
methodology

Fatigue life
prediction

Delamination
Damage criticality

Diameter
Hole
Diameter
Impact

Equivalent damage
model

D = --- ab
4

Figure 13 Schematic flow chart for assessing in-service damage.

In paper F, three different ways to characterise impact damage are suggested, see Fig. 14,
and compared to two residual strength prediction methods. These are a delamination buckling
theory and the so called soft inclusion theory and they are applied to the strength of existing
impact damaged panels. The importance of artificial delaminations are also discussed.

17

T1

(a)

Each delamination have the same

size and shape

Observed damage

Equivalent damage
45o

T2
(b)

2a

Relative
angle:
45o

2b=

0.76

Real delamination

2a

Assumed delamination

Relative angle:
90o

2b=a

Figure 14 Characterisation of damage, technique 1 and 2.

5.1

Equivalent damage methodology

A number of methods for different types of damage is suggested in Ref. [114]. An equivalent
damage size concept was introduced in Ref. [76], treating such different topics as delamination,
impact and ordinary holes. The residual strength model was formulated based on the equivalent
damage concept and a fracture criterion, i.e a model assuming macroscopically homogenous
behaviour.
Different approaches have been adopted to quantify the damaged area, which in most
cases is the delaminated region due to impact. In Ref. [115] it is postulated that the ratio of the
damaged area to some reference area is proportional to the ratio of the compliances of the plate
material and the average inclusion material. The model uses experimentally measured dimensions of the damaged region to reduce the effective average stiffness of the material in that
region and to establish the size and shape of the affected area. The nature of the damage does
not matter except that it causes changes in the stiffness in the damaged region. General aspects
regarding shape and size of individual delaminations in impact damaged structures can be
found in e.g Refs. [116-118].
In Ref. [119] a model is presented for predicting the effective anisotropic modulii of a
microcracked damaged material. By taking into account the energy dissipated through the
opening of microcracks the model can predict the effective response of the material. The total
18

strain is composed of the strain in the base material and the change in size and shape of the void
volume (microcracks). An alternative is to use the micromehanical model introduced in Ref.
[120] to study the effect of broken fibres, cracked matrix and fiber-matrix debonding. The
model is based on an integrated micro-mechanical and macromechanical fracture criterion
(IMMFC). Other micro-mechanical models can be found in e.g Ref. [121] where a repeating
cell philosophy is used. It is stated that the model can calculate elastic or non-elastic effective
modulii of composites with or without including micro-mechanical damage.
A quantitative index to describe different damages due to low-velocity damage is
described in Ref. [122]. The relations between damage and modulus degradation of the laminate, and between the damage and buckling of the laminate are preliminary given. Different
impact damages (low-velocity) are divided into four classes: slight, small, medium and larger
corresponding to different damage parameters. A more direct approach was used in Ref. [123]
where the impact damage zone was varied from 0%, which is similar to a hole, to 100% which
is no damage of the virgin material properties. Poissons ratio is assumed to remain unchanged
for the damaged area.
An analytical method has been suggested in Ref. [124] for the prediction of the compressive strength of composite laminates containing impact damage. The impact damage is characterised by ultrasonic C-scanning and the damage is modelled by an elliptical soft inclusion. The
technique employed for stress and strain calculation is Lekhniskiis complex variable technique. The stiffness of the inclusion is quantified by sublaminate buckling analysis.
Lekhnitskiis complex variable technique is an elegant technique for rapid assessment of
stress and strain of notched composite structures, Ref. [16]. This technique makes it possible to
calculate not only notched plates but also plates with regions comprising different elastic moduli, so called soft inclusions. Consider an infinite plate during general in-plane loading, see Fig.
15.

q
t
y

p
x

Figure 15 General in-plane loading for a plate with a soft inclusion.

19

Using the complex mapping technique by Lekhnitskii, stresses around and in the inclusion can
be obtained. As an alternative the delamination buckling theory, with more or less sophisticated
assumptions on blister displacement, or a combination of both, Ref. [124], can be used for
determination of residual strength.

5.2

Delamination buckling theory

Calculation methods that are cheap, fast, accurate and provides an opportunity for parametric
studies are most desirable from an engineering point of view. Semi-analytical methods based
on e.g a Rayleigh-Ritz approach are just that kind of methods. Based on the pioneering work in
Ref. [125], the theory for delamination buckling was incorporated into a computer code (SUBLAM) for damage tolerance analysis in Ref. [126]. Loads and stacking sequence information
are taken from the load distribution analysis. A conservatively assumed flaw size associated
with visible impact damage is selected. Through differentiation of the total potential energy
with respect to the minor and major axes of the ellipse the total strain energy release rate, G,
can be calculated and compared to experimental data. The authors also used the theory in the
design and development stage for quality control as a accept/reject criteria for manufactured
parts. In a recent aircraft program all composite parts were zoned for maximum allowable
delamination sizes at ultimate load. Flaw depth was not specified in the criteria, only flaw area
for the sake of conservatism, inspectability and brevity. It was assumed that the flaw was located
at the worst position (critical interface). Various aspects on the delamination buckling theory,
could be found in e.g. Refs. [125, 127, 128].
For the current approach an elliptical delamination is assumed to be present at some
depth, i.e oriented somewhere in the stacking sequence and an admissible displacement solution is assumed that fulfils the boundary conditions. Orientation and size are optional with
respect to loading directions. The problem is solved in a local coordinate system, x,y-system,
and transferred back to a user defined system, see Fig. 16. The solution is assumed to be valid
if the so called thin film approach is present, i.e sublaminate is approximately one tenth of the
total thickness.

20

y
x
0.5

Nxy

0.4

Nx

Yref

Ny

0.3

Nxy

0.2

0.1

0
-0.1
10

Xref

10

5
5

Nx

-5

-5
-10

-10

Nxy

Nxy
Ny

Delaminated
region
Nxd
Nx
Base region

Nxb

Figure 16 Problem statement.

An approximative way to calculate the strain-energy release rate for the delamination could be
found in Ref. [125].

5.3

Artificial delaminations

When artificial flaws are considered it is of vital importance to know the extent of the damage
but also to know where to locate the damage. What is lacking today is an understanding of
where the initial delamination(s) will occur, i.e a position criterion.
Artificial delaminations, located at the right place in structural verification testing and
coupled to a proper theory, could be an interesting way to characterise the strength of e.g.
impact damaged structures or to demonstrate damage tolerance. An extensive work on compression loaded plates with artificial delaminations has been performed within the GARTEUR
Action Group (AG16) on Damage Propagation in Composite Structural Elements, Ref.
[129].
Depending on size, shape and depth a delamination can first buckle in a stable manner
before it becomes critical or it can fail directly in an unstable manner which is the case for deeplying and/or small delaminations. If subjected to successive loading, the critical buckling load
will decrease. A probable cause is that adhesive forces between base-laminate and artificial
delamination will relax. The question is if artificial delaminations reflect a real behaviour or
not. The problem is of course closely related to the on-going discussions of the size and thickness of the insert for DCB and ENF-specimens.
21

SUMMARY OF APPENDED PAPERS

PAPER A
The static and cyclic fracture mechanical properties for the DCB-specimen are presented and
discussed. Numerical and experimental results are compared to standard data collection methods but also to state-of-the-art finite element analysis (FEA). The investigation covers several
materials, IM7/8552, HTA7/6376 and T300/914, and different interfaces, 0o/0o, 45o/45o and
90o/90o. It turned out that the material systems IM7/8552 and HTA7/6376 have similar critical
energy release rate whereas T300/914 has a significantly lower critical energy release rate. For
non-zero interfaces intralaminar growth occurred through the 90o or 45o plies, giving raise to a
hardening R-curve behaviour and a wavy appearance of the fracture surface. Numerically predicted curved crack front for zero-interface could be verified experimentally. The cyclic delamination growth rates increased rapidly with increased G. Comparison between specimens
with different layups showed that the specimens with a 90o/90o interface had the flattest energy
release rate distribution along the crack tip and the specimens with a 45o/45o interface had the
most uneven one. The calculated compliance compared well with the measured one after the
curved crack front had been accounted for. Energy release rate calculated with a modified beam
model has a smaller error than that calculated with the simple beam model.
PAPER B
An extensive numerical and experimental investigation regarding fracture mechanical properties for ENF-specimens is presented. For different interfaces, 0o/0o, 45o/45o and 90o/90o and
for several graphite/epoxy composites, IM7/8552, HTA7/6376 and T300/914, static and cyclic
properties are presented. State of the art numerical analyses are described and experimental
results are interpreted in terms of usefulness for applications to real structures. The FEM results
are compared with beam models and it is concluded that for the specimen with a zero interface
the beam models were acceptable whereas for the specimens with non-zero interface they did
show a large error. This could be compensated for by taking the non-homogenous nature of the
beams into account. Numerically, it was found that the effect of friction lowers the energy
release rate with a few percent. Experimental results showed that for specimens with non-zero
interfaces the delamination crack changed interface.
PAPER C
Interlaminar stress calculations are presented for straight and curved free edges and for several
stacking sequences. The interlaminar stresses are calculated using a three-dimensional finiteelement code with error control. For straight free edges the results are compared to some existing semi-analytical models. A quadratic failure criterion for prediction of initial load for delamination and the most likely position for the delamination are also discussed. It can be
concluded that interlaminar stresses exhibit a thickness effect for repeating bundles and that the
dominant component for curved free edges are the high interlaminar shear stresses and not the
normal stress. The singularity of the interlaminar components is concentrated to a narrow
region, approximately 1-2 fiber diameters in thickness and radial direction and it is concluded
that since the interlaminar stresses are so localised, a continuum approach will not provide use22

ful solutions in this area. The investigation also shows that the number of 45/0 or 45/90
alterations should be kept to a minimum to minimise the interlaminar stresses and possible
damage initiation points. Finally, it is demonstrated that the investigated semi-analytical models
do not correctly describe the interlaminar stress distribution.
PAPER D
A simplified methodology for composite life prediction is presented. The methodology is based
on fatigue failure functions which are established on a laminate level and thus the calculated
uniaxial properties, in this case with the Tsai-Hill failure criterion, will reflect the overall behaviour of the laminate. To establish the fatigue failure functions, a documented/reported SN-curve
for a multi-directional laminate is considered as a series of failure points where the failure in
the short life-region essentially is assumed to be fibre fracture and in the long life-region is
assumed to be controlled by matrix properties. For fibre failure it is also assumed that once
essential fibre failure starts to occur the remaining time to total failure is very short. The difference between first and last ply failure, in a fatigue sense, is of less importance. Once ply failure
data are established they are used in a semi-log representation to calculate the fatigue life which
is a straight forward matter since normalised, calculated, ply failure data collapse into one
curve. Different stress ratios (R=min/max) are presently handled with the Goodman-correction approach. For combined loading a series of failure envelopes, representing different times
in a log scale constitutes the behaviour.
PAPER E
An experimental investigation regarding load-sequence effects on the fatigue life of composite
structures is presented. Different elimination levels were considered for various load spectra
and for constant amplitude block loading. The spectra considered were an early design spectrum for the Gripen fighter aircraft, spectra associated with the aft fitting of the fin and to the
upper fitting of the wing, and the compression dominated spectrum Short Inverted Falstaff.
Both experimental and calculated results show, that the elimination level can be set to approximately 50% of the maximum range occurring in the load sequence. This means that considerable time and cost reductions in structural verification testing can be achieved. This will also
have impact on a life prediction methodology since only a characteristic number of loading
states needs to be considered. The resulting life for block testing, especially for a load ratio of
R=-1, is highly influenced by elimination of high load ranges in the sequence. Spectrum fatigue
test results accounting for ranges up to 90% of maximum range, the remaining load states eliminated, can be mapped on constant amplitude data. At chosen elimination levels approximately
80 to 90% of loading states were eliminated. Calculation results with the same technique, i.e
accounting for a characteristic number of cycles, show promising results.
PAPER F
In this paper two residual strength models are presented, the so called soft inclusion and delamination buckling theories, and compared to experimental results on impact damaged composite structures. Those experiments span a variety of impact events, from 8J to 55J and different
layups. The investigation has been supported by FE-technique for determination of the stress
distribution in the buckled state and for characterisation of the damaged region. It can be con23

cluded that for low-energy impact, although with conservative assumptions on stiffness reduction, that the soft inclusion theory is un-conservative for residual strength prediction. In contrast
the delamination buckling theory shows good agreement for various impact energy levels,
thicknesses and layups. The importance of repeated loading for composite structures with artificial delaminations is also demonstrated. If repeatedly loaded, the local buckling strain is
strongly reduced as compared to a non-repeated loaded structure. It can also be shown that artificial delaminations, although deep-lying, can grow in a stable manner if pre-buckled.

24

DIVISION OF WORK BETWEEN AUTHORS

PAPER A:

A Numerical and Experimental Investigation of Delamination Behaviour

in the DCB-specimen
by J. Schn, T. Nyman, A. Blom and H. Ansell

T. Nyman outlined the initial work, made the initial calculations and was also responsible for
the experimental investigation which was conducted in cooperation with BAe. J. Schn made
all the subsequent calculations and also most of the fractographic investigations. H. Ansell
made the crack growth calculations and contributed to the damage tolerance aspects of this
paper. The paper was written by T. Nyman and J. Schn under the supervision of A. Blom who
also helped with interpretation of the results.
PAPER B:

A Numerical and Experimental Investigation of a Composite ENF-specimen

by J. Schn, T. Nyman, A. Blom and H. Ansell

Same division as in paper A.

PAPER C:

Interlaminar Stresses in Composite Notched and Unnotched Laminates

by T. Nyman and M. Friberg

T. Nyman outlined the work and made the initial calculations. M. Friberg made all the subsequent calculations under the guidance of T. Nyman. The paper was written by T. Nyman with
the assistance of M. Friberg.
PAPER E:

Effects of Truncation and Elimination on Composite Fatigue Life

by T. Nyman, A. Blom and H. Ansell

T. Nyman initiated the experimental work. The theory was jointly developed by T. Nyman, A.
Blom and H. Ansell. T. Nyman and A. Blom wrote the paper.
PAPER F:

Equivalent Damage and Residual Strength for Impact Damaged

Composite Structures
by T. Nyman. A. Bredberg and J. Schn

The basic theories were developed by T. Nyman although A. Bredberg introduced the damage
characterisation methods. The experimental work was conducted by J. Schn. T. Nyman wrote
the first draft of the paper and finalized it together with J. Schn.

25

RECOMMENDATIONS FOR FUTURE WORK

This work has dealt with some basic research topics of which some are of immediate industrial
importance. Below follow a number of points for possible research in the future.
o Different threats and the associated damage need to be investigated,
especially in terms of size and position of the damage.
o The difference between structural and coupon level on residual
strength and life, i.e the scale-up effect, needs to be quantified to higher degree.
o Truncation and elimination criteria must be further looked into, especially the sorting principles for the load sequence leading to e.g. constant-amplitude events.
o Investigate the fatigue behaviour of impact damaged panels, "Fatigue
after Impact" (FAI), under realistic threats and loading conditions, including both constant-amplitude and spectrum loading
o Multiaxial static and fatigue loading of laminates with different cutouts, impact damaged and bolted joints.
o Basic fracture mechanical properties, both static and cyclic, needs further penetration. Modus III testing should be increased and interaction
formulas established. Threshold values are also of great interest.
o Simplified methods or forms for screening purposes, i.e engineering
tools with a relatively high accuracy, should be sought for residual
strength, life prediction and for interlaminar stress calculations.
o Environmental effects on static and fatigue loading should be further
investigated.

26

REFERENCES
[1]

Woodward M. R., Owens S. D., Law G. E. and Mignery, L. A., Applications of a Damage Tolerance Methodology in Aircraft Design and Production, NASA-CR-198722.
DOT/FAA/CT-92-25-Vol2 DOD/NASA/FAA Conf. on Fibrous Composite in Structural
Design, Lake Tahoe, Nov. 1991. Vol. 2, p. 1071-1081.

[2]

Whitehead R. S., Lessons Learned for Composite Structures, The first NASA
Advanced Composites Technology Conf. Seattle, Oct. 29-Nov.1, 1990, Part 1, p. 399415.

[3]

Pariraudeau R. D., and Simpson A. I., A Design Philosophy for Composite Materials,
Longman Scientific & Technical, Composite Materials in Aircraft Structures, Chapter 7
Ed. Middelton D. H.

[4]

Whitehead R.S., Kan H.P. and Saether E.S., Certification Testing Methodology for
Composite Structure -Volume I-Data Analysis, NADC-87042-60, DOT/FAA/CT-86/39,
October 1986.

[5]

Whitehead R.S., Kan H.P. and Saether E.S., Certification Testing Methodology for
Composite Structure -Volume II-Methodology Development, NADC-87042-60, DOT/
FAA/CT-86/39, October 1986.

[6]

Thionnet A. and Renard J., An Evolution Equation for Transverse Cracking in Laminated Composites Under Fatigue Loading, Rech. Aerosp. n0 1992-1.

[7]

Allen D. H., Lo D. C., Georgiou I. T. and Harris C. E., A Model for Predicting Damage
Induced Fatigue Life of Laminated Composite Structural Components, ICAS Congress,
17th, Stockholm, Sweden, Sept 9-14, 1990, Proceedings, Vol. 1, pp. 682-692.

[8]

Wang S. S., Suemasu H. and Chim S. M., Analysis of Fatigue Damage Evolution and
Associated Anisotropic Elastic Property Degradation in Random Short Fiber Composite, Journal of Composite Materials, Vol. 21 (1987), pp. 1084-1105.

[9]

Charewicz, A. and Daniel, I. M., Damage Mechanisms and Accumulation in Graphite/

Epoxy Laminates, Composite Materials: Fatigue and Fracture, ASTM STP 907, H.
T.Hahn, Ed., American Society for Testing and Materials, Philadelphia, 1986, pp. 274297.

[10] Plumtree A. and Shen G., Fatigue Damage Evolution and Life Prediction, Composites:
Proc. of the 8th Inter. Conf. on Composite Materials (ICCM/8), Honolulu, 1991, pp. 38M-10.
[11] Ratwani, M. M.and Kan, H. P., Effect of Stacking Sequence on Damage Propagation and
Failure Modes in Composite Laminates, Damage in Composite Materials, ASTM STP
775, K. L. Reifsnider, Ed., American Society for Testing and Materials,1982, pp. 211228.
[12] La Verne Leonard, Compression test results a tough nut to crack, Advanced Compos-

27

ites July/August 1989 pp 57-63.

[13] Highsmith, A. L., Stinchcomb, W. W., and Reifsnider, K. L., Effect of Fatigue-Induced
Defects on the Residual Response of Composite Materials, Effects of Defects in Composite Materials, ASTM STP 836, American Society for Testing of Materials, 1984, pp.
194-216.
[14] Soutis C., Fleck N. A. and Smith P.A., Compression Fatigue Behaviour of Notched Carbon Fibre-epoxy Laminates, International Journal of Fatigue, Vol. 13 No. 4 (1991) pp
303-312.
[15] Soutis C., Fleck N. A., Failure Prediction Technique for Compression Loaded Carbon
Fibre-Epoxy Laminate with Open Holes, Jornal of Composite Materials, Vol. 25November 1991. p 1476-1498.
[16] Lekhnitskii S. G., Anisotropic Plates 2 Ed.", Ed. by S. W. Tsai and T. Cheron, Gordon
and Breach Science Publishers, 1968.
[17] Grdal Z. and Haftka R. T., Compressive Failure Model for Anisotropic Plates with a
Cutout, AIAA JOURNAL, Vol. 25, No. 11,. Paper 86-1017 At the AIAA/ASME/ASCE/
AHS 27th Structures, Structural Dynamic and Materials Conference, San Antonio, TX,
May 19-21, 1986.
[18] Pagano, N. J., The Influence of Stacking Sequence on Laminate Strength", Journal of
Composite Materials, Vol. 5, 1971, pp 50-58.
[19] Krishna Murty A. V. and Kumar Hari H. K., "Modelling of Symmetric Laminates under
Extension", Composites Structures, Vol 11, Elsevier Pubs., pp 15-32, 1989.
[20] Icardi U. and Manuello A., "Some Considerations Regarding the Edge Stress Problem of
Laminates By Finite Elements", Computers and Structures Vol. 40, No. 3, pp. 581-597,
1991.
[21] Zhao H. and Wei J., "Three-Dimensional Finite Element Analysis on Interlaminar
Stresses of Symmetric Laminates", Computers and Structures Vol. 41. No. 4, pp 561-567,
1991.
[22] Raju, I. S. and Crews JR., H., "Interlaminar Stress Singularities at a Straight Free Edge in
Composite Laminates.", Computers and Structures, Vol 14, No. 1-2, pp21-28, 1981.
[23] Destuynder P. and Ousset Y., "Free Edge Singularity Computation ", In: ECCM-3 European Conference on Composite Materials, 3rd, Bordeaux, Mar. 20-23, 1989. London,
Elsevier, 1989, pp. 725-731.
[24] Chang, C.-C., Sandhu, R. S., Sierakowski, R. L., and Wolfe, W. E., Continuous Strain
Finite-Element Analysis of Free-Edge Effect in Laminated Composite Specimens", Journal of Composites Technology and Research, Vol. 10, No. 2, Summer 1988, pp. 54-64.
Composites Structures, Vol. 29, No. 5, pp. 783-793, 1988.
[25] Allix O. and Ladeve'ze, "Interlaminar Interface Modelling for the Prediction of Delamination.", Composite Structures 0263-8223/92, 1992 Elsevier Science Publishers Ltd.
28

England.
[26] Lucking W. M., Hoa S. V. and Sankar T. S., "A Boundary Collaction Method for Stress
Analysis of Laminate Edges.", Computers and Structures Vol. 34 No. 4, pp. 655-661,
1990.
[27] Chen J. K. and Sun C. T., " Nonlinear Analysis of Interlaminar Stress in Graphite/Epoxy
Laminates with and without Delamination Cracks", Composite Structures 8 (1987) pp.
271-285.
[28] Kassapoglo C. and Lagace P.A. "An Efficient Method for the Calculation of Interlaminar
Stresses in Composite Materials", Journal of Applied Mechanics Vol. 53, Transactions of
the ASME, December 1986, pp 744-750.
[29] Kassopoglou C., "The Effect of Combined Loading and Stacking Sequence on the Interlaminar Stress Field at Free Edges of Composite Laminates", In: American Institute of
Aeronautics and Astronautics, AIAA-89-1399-CP, pp. 2089-2097.
[30] Kassopoglou C., "Determination of Interlaminar Stresses in Composite Laminates under
Combined Loads", Journal of Reinforced Plastics and Composites, Vol. 9, January 1990
[31] Kassopoglou C. and Lagace P. A., "Closed Form Solutions for the Interlaminar Stress
Field in Angle-Ply and Cross-Ply Laminates", Journal of Composite Materials Vol. 21
No. 4/April 1987, pp 292-308.
[32] Lagace P. A., Kassapoglou C. and Brewer J. C., An Efficient Method for the Calculation
of Interlaminar Stresses in Composite Materials Due to Thermal and Mechanical
Effects., International Symposium on Composite Materials and Structures, Beijing (Peoples Republic of China), June 10-13, 1986, pp 777-783.
[33] Lagace P. A. and Saeger K. J., "Methodology for Prediction of Delamination Initiation at
Holes in Composite Laminates", In: Structures, Structural Dynamics and Materials Conference, 31st, Long Beach, CA, Apr. 2-4, 1990, AIAA-90-1019-CP, pp. 1183-1190.
[34] Yang Bing-Xian and Ye Lin, "A Boundary Layer Approach to Interlaminar Stresses in
Composite Laminates with Curved Edges", Journal of Reinforced Plastics and Composites, Vol. 7, March 1988.
[35] Folias E. S., "On the Interlaminar Stresses of a Composite Plate around the Neighborhood
of a Hole", International Journal of Solids and Structures, Vol. 25. No. 10, pp. 1193-1200,
1989.
[36] Ko Chu-Cheng and Lin Chien-Chang, "Method for Calculating the Interlaminar Stresses
in Symmetric Laminates Containing a Circular Hole", In: American Institute of Aeronautics and Astronautics, AIAA-91-0957-CP, pp. 963-973.
[37] Tang S., "A Variational Approach of Circular Cutouts in Composites", In: American
Institute of Aeronautics and Astronautics, AIAA-79-0802, pp. 326-332.
[38] Barboni R., Gaudenzi P. and Carlini S., "A Three-Dimensional Analysis of Edge Effects
in Composite Laminates with Circular Holes", Composite Structures Vol. 15, 1990, pp.
29

115-136.
[39] Brewer J.C. and Lagace P.A., "Quadratic Stress Criterion for Initiation of Delamination",
Journal of Composite Materials, Vol. 22 Nr. 12 December 1988.
[40] Kim R. Y., and Soni S. R., "Experimental and Analytical Studies On the Onset of Delamination in Laminated Composites.", Journal of Composites Material, Vol. 18 Jan. 1984,
pp. 70-80.
[41] O'Brien K. T., Characterisation of Delamination and Growth in a Composite Laminate",
Damage in Composite Materials, ASTM STP 775, American Society for Testing and
Materials, 1982.
[42] Lagace P. A., "Delamination in Composites is Toughness the Key", Sampe Journal
November/December 1986.
[43] Sun, C. T. and Zhou, S. G., "Failure Analysis of Composite Laminates with Free Edge,"
Journal of Composites Technology and Research, JCTRER, Vol. 12, No. 2, Summer
1990, pp. 91-97.
[44] Lee S. M., "Failure Mechanisms of Edge Delamination of Composites", Journal of Composite Materials, Vol. 24, November 1990.
[45] ESDU, Engineering Science Data, "Delamination and Free Edge Stresses in Laminated
Plates subjected to constant Uniaxial Strain 90021", Engineering Science Data, Structures Vol. 15.
[46] Herakovich, C. T., "Free Edge Effects in Laminated Composites", Handbook of Composites, Elsevier Science Publishers B.V., Vol 2, Chapter 4, 1989.
[47] Curtis P. T., The Fatigue Behaviour of Fibrous Composite Materials, Journal of Strain
Analysis, Vol. 24 No.4, 1989.
[48] OBrien T. K. and Martin R. H., "Round Robin Testing for Mode I Interlaminar Fracture
Toughness of Composite Materials," Journal of Composites Technology & Research,
JCTRER, Vol. 15, No. 4, Winter 1993, pp. 269-281.
[49] Ireman T. et al, "Damage propagation in composite structural elements-coupon experiments and analyses", Composite Structures 36 (1996) 209-220.
[50] Kalbermatten T. De, Jggi R., Fleler, Kausch H. H. and Davies P., "Microfocus radiography studies during mode I interlaminar fracture tests on composites", J. Materials Science Letters, Vol. 11 (1992), pp. 543-546.
[51] Polaha J.J., "Effect of Interfacial Ply Orientation on the Fracture Toughness of a Laminated Composite", AIAA-94-1537-CP, pp. 1707-1716.
[52] Olsson R., "A simplified improved beam analysis of the DCB-specimen", Composite Science and Technology, Vol. 43 (1992) pp. 329-338.
[53] Sun C. T. and Pandey R. K., "Improved method for Calculating Strain Energy Release
Rate Based on Beam Theory", AIAA journal, Vol. 32 (1994), pp. 184-189.
30

[54] Hashemi S., Kinloch A. J., and Williams J. G., Proc. R. Soc. Lond. A, Vol. 427 (1990),
pp. 451-459.
[55] Davidson B. D. and Schapery R. A., "Effect of finite width on deflection and energy releas
rate on an orthotropic double cantilever specimen", Journal of Composite Material, Vol.
22 (1988), pp. 640-656.
[56] Davidson B. D, Krger R., and Knig M. "Effect of Stacking Sequence on Energy
Release Rate Distributions in Multidirectional DCB and ENF Specimens", Engineering
Fracture Mechanics, Vol. 55 (1996), pp. 557-569.
[57] Davidson B. D., " An Analytical Investigation of Delamination Front Curvature in Double Cantilever Beam Specimens", Journal of Composite Materials, Vol. 24 (1990), pp.
1124-1137.
[58] Nilsson K. F., "On Growth of Crack Fronts in the DCB-test", Composites Engineering,
Vol.3 pp. 527-546.
[59] B. D. Davidson, and R. A. Schapery, Effect of finite width on deflection and energy
release rate of an orthotropic double cantilever specimen, J. Composite Materials, Vol.
22 (1988), pp. 640-656.
[60] B. D. Davidson, C. S. Altonen, J. J. Polaha, Effect of stacking sequence on delamination
toughness and delamination growth behavior in composite end-notched flexure specimens, Composite Materials: Testing and Design, ASTM STP 1274, R. B. Deo and C. R.
saff, Eds., 1996, pp. 393-413.
[61] Polaha J.J., Davidson B. D., Hudson R. C. and Pieracci A., " Effect of Mode Ratio, Ply
Orientation and Precracking on the Delamination Toughness of a Laminated Composite",
J. Reinforced Plastics and Composites, Vol. 15 (1996), pp.141-173.
[62] J. W. Gillespie, Jr., L. A. Carlsson, R. B. Pipes, R. Rothschilds, B. Trethewey, and A. Smiley, " Delamination growth in composite materials ", NASA CR-178066 (1986).
[63] Carlsson L. A., Gillespie Jr, and Pipes R. B., "On the Analysis and Design of the End
Notched Flexure (ENF) Specimen for Mode II Testing", J. Composite Materials, Vol. 20
(1986), pp. 594-604.
[64] Mall S., and Kochhar N. K., "Finite-Element Analysis of End-Notch Flexure Specimens",
Journal of Composites Technology & Research, Vol. 8 (1986), pp. 54-57.
[65] Russell A. J., "On the Measurement of Mode II Interlaminar Fracture Energies", DREP
Materials Report 82-0, Defence Research Establishment Pacific, Victoria, B.C., Canada,
1982.
[66] Hwu C., Kao C. J., and Chang L. E., "Delamination Fracture Criteria for Composite Laminates", J. Composite Materials, Vol. 29 (1995), pp. 1962-1987.
[67] Singh S. and Patridge I. K., Mixed-mode fracture in an interleaved carbon-fibre/epoxy
composite , Composite Science and Technology, Vol. 55 (1995), pp. 319-327.

31

[68] Corleto C. R. , and Bradley W. L., Mode II delamination fracture toughness of unidirectional graphite/epoxy composites , Composite Materials: Fatigue and Fracture, Second
Volume, ASTM STP 1012, Paul A. Lagace, Ed., 1989, pp. 201-221.
[69] Schn J. and Andersson B., "A theoretical study of the DCB and ENF test specimens", to
be submitted to Journal of Applied Mechanics.
[70] Davidson B. D., Altonen C. S. and Polaha J., Effect of Stacking Sequence on Delamination Toughness and Delamination Growth Behavior in Composite End-Notched Flexure
Specimens, Composite Materials: Testing and Design (Twelfth Volume), ASTM STP
1274, R. B. Deo and C. R. Saff, Eds., American Society for Testing and Materials, 1996,
pp. 393-413.
[71] Polaha J., Davidson B. D, Hudson R. C. and Pieracci A., Effects of Mode Ratio, Ply orientation and Precracking on the Delamination Toughness of a Laminated Composite,
Journal of Reinforced Plastics and Composites, Vol. 15 February 1996.
[72] B. D. Davidson, R. Krger, and M. Knig, Three dimensional analysis and resulting
design recommendations for unidirectional and multidirectional end-notched flexure
tests, J. Composite Materials, Vol. 29 (1995), pp. 2108-2133.
[73] Schtz D., J. J. Gerharz, Fatigue Strength of a FibreReinforced Material, Fatigue of
Fibre Reinforced Plastic Components, SEE, Society of Environmental Engineers Fatigue
Group, The City University London, 1977, pp. 82-98.
[74] Wright B. D., Fatigue Life Prediction for Carbon-Epoxy Composite Design, in Proceedings of the 16th Symposium of the International Comittee on Aeronautical Fatigue,
22-23 May 1991, Tokyo, Japan.
[75] Nyman T., Fatigue Life Prediction for Composite Materials. A literature overview,
Saab Report, 1992, TUKH R-3954.
[76] Chu C. S.,Simple and Generalised Residual Strength Model for Damaged Composites,
Journal of Aerospace Engineering, Vol.1, (1988), pp.131-141.
[77] Jen M. -H.R., Kan Y. S. and Wu I. C., Fatigue Damage in a Centrally Notched Composite Laminate due to a Two-Step Spectrum Loading, Fatigue 1994, Vol. 16 April.
[78] Charewicz, A. and Daniel, I. M., Damage Mechanisms and Accumulation in Graphite/
Epoxy Laminates, Composite Materials: Fatigue and Fracture, ASTM STP 907, H. T.
Hahn, Ed. American Society for Testing and Materials, Philadelphia, 1986, pp. 274-297.
[79] Jen M. H. R., Hsu J. M., and Lee C. H., Fatigue Damage in Centrally Notched GR/EP
Laminates, Experimental Mechanics 30 (1990) Dec. pp. 360-366.
[80] O Brien T. K.,Rigamonti M. and Zanotti C., Tension Fatigue Analysis and Life Prediction for Composite Laminates, International Journal of Fatigue 11 No. 6 (1989) pp 379393.
[81] Reifsnider, K. L. and Stinccomb, W. W., A Critical Element Model of Residual Strength
and Life of Fatigue -Loaded Composite Coupons, Composite Materials: Fatigue and
32

Fracture, ASTM STP 907, H. T. Hahn, Ed. American Society for Testing and Materials,
Philadelphia, 1986, pp. 298-313.
[82] Jen M. H. and Hsu J. M., Residual Strength in Notched Composite Materials, International Journal of Fatigue, Vol. 12 No. 4 (1990) pp. 267-274.
[83] Rotem A., Residual Strength after Fatigue Loading, International Journal of Fatigue
No. 10, January 1988, pp 27-31.
[84] Sims, D. F. and Brogdon, Criterion for Fiber Reinforced Materials, Journal of Composite Materials, Vol, V. H.,Fatigue Behaviour of Composites Under Different Loading
Modes, Fatigue of Filamentary Composite Materials, ASTM STP 636, K. L. Reifsnider
and K. N. Lauraitis, Eds., American Society for Testing and Materials, 1977 pp. 185 205.
[85] Hashin, Z and Rotem A., A Fatigue Failure Criterion for Fiber Reinforced Materials, J.
Composite Mater, Vol. 7 (October 1973), p.448
[86] Schtz D., J. J. Gerharz, and Alschweig, Fatigue Properties of Unnotched, Notched, and
Jointed Specimens of Graphite/Epoxy Composite, Fatigue of Fibrous Composite Materials, ASTM STP 723, American Society for Testing and Materials, 1981, pp. 31 -47.
[87] Fawaz, Z. and Ellyin, F, Fatigue Failure Model for Fibre-Reinforced Materials under
General Loading Conditions, Journal of Composite Materials, Vol. 28, No.15/1994.
[88] Fawaz, Z. and Ellyin, F., A New Methodology for the Prediction of Fatigue Failure in
Multidirectional Fiber-Reinforced Laminates, Composite Structures and Technology 53
(1995) 47-55.
[89] Ellyin F. and El-Kadi, AFatigue Failure Criterion for Fiber Reinforced Composite
Laminate, Composite Structures, Vol. 15 (1990), pp. 61-74.
[90] Rotem A., The Fatigue Behaviour of Composite Laminates under Various Mean
Stresses, Composites Structures 17 (1991), pp 113-126.
[91] Rotem A., The Fatigue Behaviour of Orthotropic Laminates under Tension-Compression Loading, International Journal of Fatigue No. 3 (1991) pp 209-215.
[92] Sun C. T. and Zheng S., "Delamination Characteristics of Double-Cantilever Beam and
End-notched Flexure Composite Specimens", Composite Science and Technology, Vol.
56 (1996), pp. 451-459.
[93] Andersson B., Falck U., Babuska I., and Von Petersdorff T., "Reliable Stress and Fracture
Mechanics Analysis of Complex Components Using a h-p Version of FEM", Int. J.
Numer. Methods. Eng., Vol.38 (1995), pp. 2135-2163.
[94] Herakovich, C. T., "Edge Effects and Delamination Failures ", Journal of Strain Analysis
Vol. 24 No. 4. 1989.
[95] Lagace P. A., Brewer J. C. and Kassapoglou C., "The Effect of Thickness on Interlaminar
Stresses and Delamination in Straight-Edged Laminates", Journal of Composites Tech33

nology and Research, Vol. 9, No. 3, Fall 1987, pp. 81-87.

[96] Hong C. S., "Suppression of Interlaminar Stresses of Thick Laminates using Sublaminate
Approach", In: 32nd International SAMPE Symposium April 6-9, 1987.
[97] OBrien T. K., Murri G. B. and Salpekar S. A., Interlaminar Shear Fracture Toughness
and Fatigue Thresholds for Composite Materials, Composite Materials: Fatigue and
Fracture, Second Volume, ASTM STP 1012, Paul A. Lagace, ED., American Society for
Testing and Materials, Philadelphia, 1989, pp. 222-250.
[98] Walter, R. W., June, R. R., and McCarty, J. E., Designing for Integrity in Long-Life
Composite Aircraft Structures,Fatigue of Filamentary Composite Materials, ASTM
STP 636,K: L. Reifsnider and K. N. Lauraitis, Eds., American Society for Testing and
Materials, 1977 pp. 228-247.
[99] Talreja R., Fatigue of Composite Materials, Tecnomic Publishing Company, Inc.,
1987.
[100] Wang et. al, Failure Analysis of Composite Laminates, TECHNOMIC Publishing
Company, Inc. (1085)
[101] Avery, J. G., Design Manual for Impact Damage Tolerant Aircraft Structure, AGARDograph No.238 (1980).
[102] Rotem A., The Anisotropic Fatigue Behaviour of Isotropic Laminates, Fatigue, 1994,
Vol 16, June, pp 266-272.
[103] Rotem, A., Effect of the frequency of the Alternating Load on Fatigue Delamination of
Laminates, SAMPE Journal Vol. 30, No. 3, May/June 1994.
[104] Masters, J. E. and Reifsnider, K. L. An Investigation of Cumulative Damage Development in Quasiisotropic Graphite/Epoxy Laminates, Damage in Composite Materials,
ASTM STP 775, K. L. Reifsnider, Ed., American Society for Testing and Materials,
1982, pp. 40-62.
[105] Gathercole N., Reiter H. and Harris B., "Life prediction for fatigue of T800/5245 carbonfibre composites:I. Constant-amplitude loading", Fatigue, 1994, Vol. 16, November, pp.
523-532.
[106] Adam T., Gathercole N., Reiter H. and Harris B., "Life prediction for fatigue of T800/
5245 carbon-fibre composites:I. Variable-amplitude loading", Fatigue, 1994, Vol. 16,
November, pp. 533-547.
[107] Harris B., Gathercole N., Reiter H. and Adam T., "Fatigue of carbon-fibre-reinforced
plastics under block-loading conditions", Composites Part A, 28A (1997) pp 327-337.
[108] Badaliance R. Dill H. D., Effects of Fighter Attack Spectrum on Composite Fatigue
Life, Air Force Wright Aeronautical Laboratories, AFAWL-TR-81-3001.
[109] Badaliance R. Dill H. D. and Potter J. M., "Effects of Spectrum Variations on Fatigue Life
of Composites", Composite Materials: Testing and Design (Sixth Conference), ASTM
34

STP 787, I. M. Daniel, Ed., American Society for Testing and Materials, 1982, p.p 274286.
[110] Sanger K. B., Dill H. D. and Kautz E. F., "Certification Testing Methodology for Fighter
Hybrid Structure", Composite Materials: Testing and Design (Ninth Volume), ASTM
STP 1059, S. P. Garbo, Ed., American Society for Testing and Materials, Philadelphia,
1990, pp. 34-47.
[111] Harlan R. A., Damage Tolerance and Durability Testing for F/A-18 E/F Composite
Materials Structures, American Institute of Aeronautics and Astronautics, AIAA-961320-CP, 1996.
[112] Phillips E. P., "Effects of Truncation of a Predominantly Compression Load Spectrum on
the Life of a Notched Graphite/Epoxy Laminate," Fatigue of Fibrous Composite Materials, ASTM STP 723, American Society for Testing and Materials, 1981, pp. 197-212.
[113] Dorey G., Damage Tolerance and Damage Assessment in Advanced Composites, Elsevier Applied Science, Advanced Composites, Ed. I. K. Partridge, Chp. 11. 1990.
[114] Chamis, C. C., and Ginty, C. A., "Fiber Composite Structural Durability and Damage Tolerence: Simplified Predictive Methods", Composite Materials: Fatigue and Fracture, Second Volume, ASTM STP 1012, Paul A. Lagace, Ed., American Society for Testing and
Materials, Philadelphia, 1989, pp. 338-355.
[115] El-Zein M. S. and Reifsnider, On the prediction of Tensile Strength after Impact of Composite Laminates, Journal of Composites Technology & Research, JCTRER, Vol. 12,
No. 3, Fall 1990, pp. 147-154.
[116] Hull D., Shi Y.B., Damage mechanism characterization in composite damage tolerance
investigations., Composite structures, 23 (1993).
[117] Beks F.A., Examination of impact response and damage of composite laminates., FFA
TN 1996-29, The Aeronautical Research Institute of Sweden, Bromma, 1996
[118] Liu D., Impact-Induced delamination-A view of bending stiffness mismatching., J. of
Composite Materials, 22 (July 1988).
[119] Santare M. H., Crocmbe A. D. and Anlas G., Anisotropic Effective Moduli with Microcracks, Engineering Fracture Mechanics Vol. 52, No. 5. pp 833-842, 1995.
[120] Mahishi J. M., An Integrated Micro mechanical and Macromechanical Approach to
Fracture Behaviour of Fiber-Reinforced Composites, Engineering Fracture Mechanics
Vol. 25, No. 2, pp. 197-228, 1986.
[121] Kwon, Y. W., Calculation of Effective Moduli of Fibrous Composites with Micromechanical Damage, Composite Structures 25 (1993) pp 187-192.
[122] Yang L. and Shuang L. Z., The Preliminary Analysis of Reliability for a Composite
Structure with Damage, Engineering Fracture Mechanics Vol. 37, No. 6, pp. 1157-1163,
1990.

35

[123] Chen V. L., Wu H.-Y and Yeh H.-Y, A Parametric Study of Residual Strength and Stiffness for Impact Damaged Composites, Composite Structures 25 (1993) pp 267-275.
[124] Xiong Y., Poon C., Straznicky P. V. and Vietinghoff H., A Prediction Method for the
Compressive Strength of Impact Damaged Composite Laminates, Composite Structures
30 (1995) p. 357-367.
[125] Chai, H. and Babcock, C. D., Two-Dimensional Modelling Compressive Failure in
Delaminated Regions, Journal of Composite Materials, Vol. 19 - January 1985.
[126] Woodward M. R., Owens S. D., Law G. E. and Mignery, L. A., Applications of a Damage Tolerance Methodology in Aircraft Design and Production, NASA-CR-198722.
DOT/FAA/CT-92-25-Vol2 DOD/NASA/FAA Conf. on Fibrous Composite in Structural
Design, Lake Tahoe, Nov. 1991. Vol. 2, p. 1071-1081.
[127] Whitcomb, J.D. and Shivakumar, K. N., Buckling of s Sublaminate in a Quasi-Isotropic
Composite Laminate, Journal of Composite Materials, Vol. 19- January 1985.
[128] Davidsson, B. D., Delamination Buckling: Theory and Experiment, Journal of Composite materials, Vol. 25 - October 1991.
[129] Ireman et.al, GARTEUR-STRUCTURES AND MATERIALS -AG 16 Damage Propagation in Composite Structural Elements Final Report Part2 Coupons, GARTEUR TP101-2, Saab FKH R-4109.

36