This action might not be possible to undo. Are you sure you want to continue?
Lecture notes  course 4A780
Concept version
Dr.ir. P.J.G. Schreurs
Eindhoven University of Technology
Department of Mechanical Engineering
Materials Technology
September 13, 2011
Contents
1 Introduction 1
2 Fracture mechanics 9
2.1 Fracture mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Shearing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 Cleavage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.3 Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.4 Crazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.5 Deadhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Ductile  brittle behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.1 Charpy vnotch test . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Theoretical strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.1 Discrepancy with experimental observations . . . . . . . . . . . . . . . 16
2.3.2 Griﬃth’s experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.3 Crack loading modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3 Experimental techniques 19
3.1 Surface cracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Electrical resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Xray . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Ultrasound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.5 Acoustic emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.6 Adhesion tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4 Fracture energy 23
4.1 Energy balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2 Griﬃth’s energy balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.3 Griﬃth stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3.1 Discrepancy with experimental observations . . . . . . . . . . . . . . . 26
4.4 Compliance change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.4.1 Fixed grips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4.2 Constant load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
I
II
5 Stress concentrations 31
5.1 Deformation and strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.3 Linear elastic material behavior . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.4 Equilibrium equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5.5 Plane stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.6 Plane strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.7 Displacement method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.8 Stress function method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.9 Circular hole in ’inﬁnite’ plate . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.10 Elliptical hole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6 Crack tip stresses 45
6.1 Complex plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.1.1 Complex variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.1.2 Complex functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.1.3 Laplace operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.1.4 Biharmonic equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.2 Solution of biharmonic equation . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.2.1 Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.2.2 Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.2.3 Choice of complex functions . . . . . . . . . . . . . . . . . . . . . . . . 50
6.2.4 Displacement components . . . . . . . . . . . . . . . . . . . . . . . . . 50
6.3 Mode I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.1 Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.2 Stress components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3.3 Stress intensity factor . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.3.4 Crack tip solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.4 Mode II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.4.1 Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.4.2 Stress intensity factor . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.4.3 Crack tip solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.5 Mode III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.5.1 Laplace equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6.5.2 Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6.5.3 Stress components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
6.5.4 Stress intensity factor . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.5.5 Crack tip solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.6 Crack tip stress (mode I, II, III) . . . . . . . . . . . . . . . . . . . . . . . . . 56
6.6.1 Kzone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.7 SIF for speciﬁed cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.8 Kbased crack growth criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6.9 Relation G−K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6.10 The critical SIF value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.10.1 K
Ic
values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
III
7 Multimode crack loading 63
7.1 Stress component transformation . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.2 Multimode load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
7.3 Crack growth direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
7.3.1 Maximum tangential stress criterion . . . . . . . . . . . . . . . . . . . 69
7.3.2 Strain energy density (SED) criterion . . . . . . . . . . . . . . . . . . 72
8 Dynamic fracture mechanics 77
8.1 Crack growth rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
8.2 Elastic wave speeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
8.3 Crack tip stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
8.3.1 Crack branching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.3.2 Fast fracture and crack arrest . . . . . . . . . . . . . . . . . . . . . . . 81
8.4 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
9 Plastic crack tip zone 83
9.1 Von Mises and Tresca yield criteria . . . . . . . . . . . . . . . . . . . . . . . . 83
9.2 Principal stresses at the crack tip . . . . . . . . . . . . . . . . . . . . . . . . . 84
9.3 Von Mises plastic zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
9.4 Tresca plastic zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
9.5 Inﬂuence of the plate thickness . . . . . . . . . . . . . . . . . . . . . . . . . . 87
9.6 Shear planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
9.7 Plastic constraint factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
9.8 Plastic zone in the crack plane . . . . . . . . . . . . . . . . . . . . . . . . . . 90
9.8.1 Irwin plastic zone correction . . . . . . . . . . . . . . . . . . . . . . . . 90
9.8.2 DugdaleBarenblatt plastic zone correction . . . . . . . . . . . . . . . 91
9.8.3 Plastic zones in the crack plane . . . . . . . . . . . . . . . . . . . . . . 92
9.9 Small Scale Yielding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
10 Nonlinear Fracture Mechanics 95
10.1 Cracktip opening displacement . . . . . . . . . . . . . . . . . . . . . . . . . . 95
10.1.1 CTOD by Irwin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
10.1.2 CTOD by Dugdale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
10.1.3 CTOD crack growth criterion . . . . . . . . . . . . . . . . . . . . . . . 98
10.2 Jintegral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
10.2.1 Integral along closed curve . . . . . . . . . . . . . . . . . . . . . . . . 99
10.2.2 Path independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
10.2.3 Relation J ∼ K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
10.3 HRR crack tip stresses and strains . . . . . . . . . . . . . . . . . . . . . . . . 101
10.3.1 RambergOsgood material law . . . . . . . . . . . . . . . . . . . . . . 102
10.3.2 HRRsolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
10.3.3 Jintegral crack growth criterion . . . . . . . . . . . . . . . . . . . . . 103
11 Numerical fracture mechanics 105
11.1 Quadratic elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
11.2 Crack tip mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
11.3 Special elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
IV
11.4 Quarter point elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
11.4.1 Onedimensional case . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
11.5 Virtual crack extension method (VCEM) . . . . . . . . . . . . . . . . . . . . . 110
11.6 Stress intensity factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
11.7 Jintegral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
11.7.1 Domain integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
11.7.2 De Lorenzi Jintegral : VCE technique . . . . . . . . . . . . . . . . . . 113
11.8 Crack growth simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
11.8.1 Node release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
11.8.2 Moving Crack Tip Mesh . . . . . . . . . . . . . . . . . . . . . . . . . . 114
11.8.3 Element splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
11.8.4 Smeared crack approach . . . . . . . . . . . . . . . . . . . . . . . . . . 116
12 Fatigue 119
12.1 Crack surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
12.2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
12.3 Fatigue load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
12.3.1 Fatigue limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
12.3.2 (SN)curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
12.3.3 Inﬂuence of average stress . . . . . . . . . . . . . . . . . . . . . . . . . 122
12.3.4 (PSN)curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
12.3.5 High/low cycle fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
12.3.6 Basquin relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
12.3.7 MansonCoﬃn relation . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
12.3.8 Total strainlife curve . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
12.4 Inﬂuence factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
12.4.1 Load spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
12.4.2 Stress concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
12.4.3 Stress gradients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
12.4.4 Material properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
12.4.5 Surface quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
12.4.6 Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
12.5 Crack growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
12.5.1 Crack growth models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
12.5.2 Paris law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
12.5.3 Fatigue life . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
12.5.4 Other crack grow laws . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
12.5.5 Crack growth at low cycle fatigue . . . . . . . . . . . . . . . . . . . . . 134
12.6 Load spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
12.6.1 Random load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
12.6.2 Tensile overload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
12.7 Design against fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
13 Engineering plastics (polymers) 143
13.1 Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
13.1.1 Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
13.1.2 Properties of engineering plastics . . . . . . . . . . . . . . . . . . . . . 144
V
13.2 Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
14 Cohesive zones 149
14.1 Cohesive zone models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
14.1.1 Polynomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
14.1.2 Piecewise linear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
14.1.3 Rigidlinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
14.1.4 Exponential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
14.2 Weighted residual formulation with cohesive zones . . . . . . . . . . . . . . . 157
14.3 Twodimensional CZ element . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
14.3.1 Local vector base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
14.3.2 Iterative procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
14.4 CZ for large deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
14.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
14.5.1 Polymer coated steel for packaging . . . . . . . . . . . . . . . . . . . . 166
14.5.2 Easy peeloﬀ lid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
14.5.3 Solder joint fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
A Laplace equation a1
B Derivatives of Airy function a3
Chapter 1
Introduction
Important aspects of technological and biological structures are stiﬀness and strength. Re
quirements on stiﬀness, being the resistance against reversible deformation, may vary over a
wide range. Strength, the resistance against irreversible deformation, is always required to
be high, because this deformation may lead to loss of functionality and even global failure.
Fig. 1.1 : Stiﬀness and strength.
Continuum mechanics
When material properties and associated mechanical variables can be assumed to be con
tinuous functions of spatial coordinates, analysis of mechanical behavior can be done with
Continuum Mechanics. This may also apply to permanent deformation, although this is as
sociated with structural changes, e.g. phase transformation, dislocation movement, molecular
slip and breaking of atomic bonds. The only requirement is that the material behavior is
studied on a scale, large enough to allow small scale discontinuities to be averaged out.
1
2
When a threedimensional continuum is subjected to external loads it will deform. The
strain components, which are derived from the displacements u
i
(i = 1, 2, 3) by diﬀerentiation
– ( )
,j
– with respect to spatial coordinates x
j
(j = 1, 2, 3), are related by the compatibility
relations.
The stress components σ
ij
(i, j = 1, 2, 3) must satisfy the equilibrium equations – partial
diﬀerential equations – and boundary conditions.
In most cases the equilibrium equations are impossible to solve without taking into
account the material behavior, which is characterized by a material model, relating stress
components σ
ij
to strain components ε
kl
(k, l = 1, 2, 3).
x
A
0
A
V
x
0
V
0
u
O
e
1
e
2
e
3
Fig. 1.2 : Deformation of a continuum.
 volume / area V
0
, V / A
0
, A
 base vectors {e
1
, e
2
, e
3
}
 position vector x
0
, x
 displacement vector u
 strains ε
kl
=
1
2
(u
k,l
+u
l,k
)
 compatibility relations
 equilibrium equations σ
ij,j
+ρq
i
= 0 ; σ
ij
= σ
ji
 density ρ
 load/mass q
i
 boundary conditions p
i
= σ
ij
n
j
 material model σ
ij
= N
ij
(ε
kl
)
Material behavior
To investigate, model and characterize the material behaviour, it is necessary to do experi
ments. The most simple experiment is the tensile/compression loading of a tensile bar.
3
Straintime and stresstime curves reveal much information about the material behavior,
which can be concluded to be elastic : reversible, timeindependent; viscoelastic : reversible,
timedependent; elastoplastic : irreversible, timeindependent; viscoplastic : irreversible,
timedependent; damage : irreversible, decreasing properties.
Fig. 1.3 : Tensile test.
t t
ε
t
2
σ
t
1
t
2
t
1
t
t
2
t
ε σ
t
1
t
2
t
1
t t
ε σ
t
1
t
2
t
1
t
2
ε
e
ε
p
t
2
t t
ε σ
t
1
t
2
t
1
Fig. 1.4 : Stress excitation and strain response as function of time.
Topleft →anticlockwise: elastic, elastoplastic, viscoelastic, viscoplastic.
Stressstrain curves may show linear, nonlinear, hardening and softening behavior.
ε ε
σ σ
σ
ε ε
σ
Fig. 1.5 : Stressstrain curves for hardening (left) and sofening (right) material behavior.
Fracture
When material damage like microcracks and voids grow in size and become localized, the
averaging procedure can no longer be applied and discontinuities must be taken into account.
This localization results in a macroscopic crack, which may grow very fast, resulting in global
failure.
4
Although early approaches have striven to predict fracture by analyzing the behavior of
atomic bonds, Griﬃth has shown in 1921 that attention should be given to the behavior of
an existing crack.
Fig. 1.6 : Tensile test with axial elongation and fracture.
Fracture mechanics
In fracture mechanics attention is basically focused on a single crack. Theoretical concepts
and experimental techniques have been and are being developed, which allow answers to
questions like:
• Will a crack grow under the given load ?
• When a crack grows, what is its speed and direction ?
• Will crack growth stop ?
• What is the residual strength of a construction (part) as a function of the (initial) crack
length and the load ?
• What is the proper inspection frequency ?
• When must the part be repaired or replaced ?
Several ﬁelds of science are involved in answering these questions : material science and
chemistry, theoretical and numerical mathematics, experimental and theoretical mechanics.
As a result, the ﬁeld of fracture mechanics can be subdivided in several specializations, each
with its own concepts, theory and terminology.
5
Fig. 1.7 : Crack in a bicycle crank.
(Source: internet)
Experimental fracture mechanics
Detection of cracks is done by experimental techniques, ranging from simple and cheap to
sophisticated and expensive. Experimental Fracture Mechanics (EFM) is about the use and
development of hardware and procedures, not only for crack detection, but, moreover, for the
accurate determination of its geometry and loading conditions.
Fig. 1.8 : Experimental tensile equipment.
(Source: Internet)
Linear elastic fracture mechanics
A large ﬁeld of fracture mechanics uses concepts and theories in which linear elastic material
behavior is an essential assumption. This is the case for Linear Elastic Fracture Mechanics
(LEFM).
Prediction of crack growth can be based on an energy balance. The Griﬃth criterion
states that ”crack growth will occur, when there is enough energy available to generate new
crack surface.” The energy release rate is an essential quantity in energy balance criteria. The
6
resulting crack growth criterion is referred to as being global, because a rather large volume
of material is considered.
The crack growth criterion can also be based on the stress state at the crack tip. This
stress ﬁeld can be determined analytically. It is characterized by the stress intensity factor.
The resulting crack growth criterion is referred to as local, because attention is focused at a
small material volume at the crack tip.
Assumption of linear elastic material behavior leads to inﬁnite stresses at the crack tip.
In reality this is obviously not possible: plastic deformation will occur in the crack tip region.
Using yield criteria (Von Mises, Tresca), the crack tip plastic zone can be determined. When
this zone is small enough (Small Scale Yielding, SSY), LEFM concepts can be used.
Fig. 1.9 : Plastic strain zone at the crack tip.
Dynamic fracture mechanics
It is important to predict whether a crack will grow or not. It is also essential to predict the
speed and direction of its growth. Theories and methods for this purpose are the subject of
Dynamic Fracture Mechanics (DFM).
Fig. 1.10 : Impact of a bullet in a plate [39].
7
Nonlinear fracture mechanics
When the plastic crack tip zone is too large, the stress and strain ﬁelds from LEFM are not
valid any more. This is also the case when the material behavior is nonlinear elastic (e.g. in
polymers and composites). Crack growth criteria can no longer be formulated with the stress
intensity factor.
In ElasticPlastic Fracture Mechanics (EPFM) or NonLinear Fracture Mechanics (NLFM)
criteria are derived, based on the Crack Tip Opening Displacement. Its calculation is possible
using models of Irwin or DugdaleBarenblatt for the crack tip zone.
Another crack growth parameter, much used in NLFM, is the Jintegral, which charac
terizes the stress/deformation state in the crack tip zone.
Numerical techniques
Analytical calculation of relevant quantities is only possible for some very simple cases. For
more practical cases numerical techniques are needed. Numerical calculations are mostly done
using the Finite Element Method (FEM). Special quarterpoint crack tip elements must be
used to get accurate results.
Fig. 1.11 : Loading and deformation of a cracked plate. (A quarter of the plate is
modelled and shown.)
Fatigue
When a crack is subjected to a timedependent load, be it harmonic or random, the crack will
gradually grow, although the load amplitude is very small. This phenomenon is called Fatigue
Crack Propagation (FCP) and may unexpectedly and suddenly result in Fatigue Failure.
8
Fig. 1.12 : Engine and crank axis.
Outline
Many books and papers on fracture mechanics and fatigue exist, a lot of which are rather
intimidating for a ﬁrst reader, because they are either too specialized and/or lack a proper
introduction of concepts and variables. This is partly due to the fact that Fracture Mechanics
has developed into many specialized subjects, all focused on diﬀerent applications. In this
course a number of these subjects are discussed, with an emphasis on a clear introduction of
concepts and derivation of variables. The main goal is to give an overview of the whole ﬁeld.
For more detail, the reader is referred to the specialized publications, some of which can be
found in the References.
Chapter 2
Fracture mechanics
In this chapter we ﬁrst discuss some mechanisms, which can be recognized after fracture
has occured. They are generally related to the amount of dissipated energy during crack
growth and therefore to ductile or brittle fracture. The theoretical strength of a material
can be predicted from the maximum bond strength between atoms. The result of these
very approximating calculations appears to be not in accordance with experimental ﬁndings.
Griﬃth was the ﬁrst to recognize that attention must be focused to imperfections like already
existing cracks. Fracture Mechanics is thus started with the experimental studies by Griﬃth
in 1921 and has since then developed in various directions.
2.1 Fracture mechanisms
The way a crack propagates through the material is indicative for the fracture mechanism.
Visual inspection of the fracture surface gives already valuable information. The next mech
anisms are generally distinguished.
• shear fracture
• cleavage fracture
• fatigue fracture
• crazing
• deadhesion
2.1.1 Shearing
When a crystalline material is loaded, dislocations will start to move through the lattice due
to local shear stresses. Also the number of dislocations will increase. Because the internal
structure is changed irreversibly, the macroscopic deformation is permanent (plastic). The
dislocations will coalesce at grain boundaries and accumulate to make a void. These voids
will grow and one or more of them will transfer in a macroscopic crack. One or more cracks
may then grow and lead to failure.
Because the origin and growth of cracks is provoked by shear stresses, this mechanism
is referred to as shearing. Plastic deformation is essential, so this mechanism will generally
be observed in FCC crystals, which have many closedpacked planes.
9
10
The fracture surface has a ’doughlike’ structure with dimples, the shape of which indicate
the loading of the crack. For some illustrative pictures, we refer to the book of Broek [10].
Fig. 2.1 : Dislocation movement and coalescence into grain boundary voids, resulting in
dimples in the crack surface [30].
2.1.2 Cleavage
When plastic deformation at the crack tip is prohibited, the crack can travel through grains by
splitting atom bonds in lattice planes. This is called intra or transgranular cleavage. When
the crack propagates along grain boundaries, it is referred to as intergranular cleavage.
This cleavage fracture will prevail in materials with little or no closedpacked planes,
having HCP or BCC crystal structure. It will also be observed when plastic deformation is
11
prohibited due to low temperature or high strain rate. As will be described later, a three
dimensional stress state may also result in this mechanism.
Intergranular cleavage will be found in materials with weak or damaged grain bound
aries. The latter can be caused by environmental inﬂuences like hydrogen or high temperature.
The crack surface has a ’shiny’ appearance. The discontinuity of the lattice orientations
in neighboring grains will lead to socalled cleavage steps, which resemble a ’river pattern’.
Nice pictures can be found in [11].
intergranulair intragranulair
Fig. 2.2 : Inter and intragranular cleavage fracture.
2.1.3 Fatigue
When a crack is subjected to cyclic loading, the crack tip will travel a very short distance in
each loading cycle, provided that the stress is high enough, but not too high to cause sudden
global fracture. With the naked eye we can see a ’clam shell’ structure in the crack surface.
Under a microscope ’striations’ can be seen, which mark the locations of the crack tip after
each individual loading cycle.
This mechanism is referred to as fatigue. Because crack propagation is very small in
each individual load cycle, a large number of cycles is needed before total failure occurs. The
number of cycles to failure N
f
is strongly related to the stress amplitude ∆σ =
1
2
(σ
max
−σ
min
)
and the average stress σ
m
=
1
2
(σ
max
+σ
min
).
Nice pictures of macroscopic and microscopic fatigue crack surfaces can again be found
in [11].
Fig. 2.3 : Clam shell fatigue crack surface and striations [30].
12
2.1.4 Crazing
In a polymer material submicrometer voids may initiate when a critical load level is ex
ceeded. Sometimes, one or a limited number of these crazes grow locally to generate a large
and fatal crack. In other circumstances the crazes spread out over a larger area. This is indi
cated as ”stress whitening”, because the crazes refract the light, resulting in a white colored
appearance. Some nice pictures of crazes can be found in [29].
Fig. 2.4 : Crazes in polystyrene on a macroscopic and microscopic scale.
(Source: Internet)
2.1.5 Deadhesion
Adhesion refers to bonding between atoms of diﬀerent materials, while cohesion refers to
bonding between atoms of one and the same material. Applications include the joining of
two diﬀerent entities such as structural parts and the adhesion of a generally thin layer on a
generally thicker substrate. In the latter case the thin layer is also often referred to as surface
layer, thin ﬁlm or coating.
Surface layers may be metallic – metals (M) and their nitrides (MNx) or oxides (MOx)
–, anorganic – oxides, carbides, diamond – or organic – polymers. Thicknesses vary from
several nanometers to hundreds of micrometers. Substrates can also be metallic – metals and
alloys –, anorganic – ceramics, glass – and organic.
The adhesion of the layer to the substrate is determined by chemical bond strength and
highly inﬂuenced by initial stresses and damage in the surface layer and the roughness of
the substrate. Initial residual stresses can be classiﬁed as thermal , due to solidiﬁcation and
cooling , and mechanical , due to volume mismatch (penetration) and plastic strain (impact).
The adhesion strength can be characterized by the maximum normal strength, the maximum
13
tangential strength or the work needed for separation. Many experimental techniques are
available to determine these parameters. Modeling of the adhesion strength is often done
with cohesive zone (CZ) models, where the mentioned parameters play an important role.
2.2 Ductile  brittle behavior
When a crack propagates, new free surface is generated, having a speciﬁc surface energy γ,
which for solid materials is typically 1 [Jm
−2
]. This energy is provided by the external load
and is also available as stored elastic energy. Not all available energy, however, is used for the
generation of new crack surfaces. It is also transformed into other energies, like kinetic energy
or dissipative heat. When a lot of available energy is used for crack growth, the fracture is
said to be brittle. When a lot of energy is transformed into other energies, mainly due to
dissipative mechanisms, the fracture is indicated to be ductile.
Although boundary and environmental conditions are of utmost importance, it is com
mon practice to say that a certain material is brittle or ductile. A ﬁrst indication of this
follows from the stressstrain curve, registered in a tensile test up to fracture of the tensile
bar. The area under the curve is a measure for the dissipated energy before failure. Tensile
curves for a variety of materials are shown in the ﬁgure.
Because dissipation is associated with plastic deformation, shear fracture is often found
in materials which show ductile fracture. When plastic deformation and thus dissipation is
less, fracture is more brittle.
10 0 100
ε (%)
σ
ABS, nylon, PC
PE, PTFE
Fig. 2.5 : Tensile test up to fracture and various stressstrain curves.
2.2.1 Charpy vnotch test
Although the tensile stressstrain curve already provides an indication for brittle/ductile fail
ure, the standard experiment to investigate this is the Charpy Vnotch test. The main
advantage of this test is that it provides a simple measure for the dissipated energy during
fast crack propagation.
The specimen is a beam with a 2 mm deep Vshaped notch, which has a 90
o
angle and
a 0.25 mm root radius. It is supported and loaded as in a threepoint bending test. The load
is provided by the impact of a weight at the end of a pendulum. A crack will start at the
tip of the Vnotch and runs through the specimen. The material deforms at a strain rate
of typically 10
3
s
−1
. The energy which is dissipated during fracture can be calculated easily
14
from the height of the pendulum weight, before and after impact. The dissipated energy is
the Impact Toughness C
v
[J].
Fig. 2.6 : Charpy Vnotch test.
The impact toughness can be determined for various specimen temperatures T. For intrinsic
brittle materials like high strength steel, the dissipated energy will be low for all T. For
intrinsic ductile materials like FCCmetals, C
v
will be high for all T. A large number of
materials show a transition from brittle to ductile fracture with increasing temperature. The
transition trajectory is characterized by three temperatures :
NDT : Nil Ductility Temperature
FATT : Fracture Appearance Transition Temperature (T
t
)
FTP : Fracture Transition Plastic
As a rule of thumb we have for BCC alloys T
t
= 0.1 `a 0.2T
m
and for ceramics T
t
= 0.5 `a 0.7T
m
,
where T
m
is the melting temperature.
More on the Charpy test can be found in [29]. Other tests to determine the brittleness
are the Izod test and Drop Weight Test.
T
C
v
low strength
bcc metals
Be, Zn, ceramics
high strength metals
Al, Ti alloys
fcc (hcp) metals
NDT FATT FTP T
C
v
T
t
Fig. 2.7 : C
v
values as a function of temperature T.
15
2.3 Theoretical strength
Metal alloys consist of many crystals, each of which has lattice planes in a certain spatial
orientation. In a ﬁrst attempt to calculate the strength of such a crystalline material, one
atomic plane perpendicular to the tensile load is considered. The bonding force f between
two atoms in two neighboring planes depends on their distance r and can be calculated from a
potential. The interaction force can be approximated by a sine function, using the equilibrium
distance a
0
, the half wavelength λ and the maximum force f
max
. With this approximation,
the interaction force is zero and the bond is broken, when r = a
0
+
1
2
λ.
The interaction force between all atoms in an area S of the two lattice planes can be
calculated by addition. The stress σ, which is the ratio of this total force and the area S, can
be expressed in the maximum stress σ
max
, being the theoretical material strength.
x
r
1
2
λ
f
a
0
r
f f
S
σ
x
Fig. 2.8 : Atomic bond strength between atom in a lattice.
f(x) = f
max
sin
_
2πx
λ
_
; x = r −a
0
σ(x) =
1
S
f(x) = σ
max
sin
_
2πx
λ
_
The theoretical strength σ
max
can be determined from an energy balance. When the interac
tion force and thus the interatomic distance increases, elastic energy is stored in each bond.
When the bond breaks at x = λ/2, the stored energy is released. The stored elastic energy
per unit of area is U
i
and can be calculated by integration.
It is assumed that all bonds in the area S snap at the same time and that all the stored
energy is transformed in surface energy, which, per unit of area is U
a
= 2γ, with γ the speciﬁc
surface energy of the material.
From the energy balance U
i
= U
a
, the wave length of the sine function can be calculated
and the stress as a function of the displacement x can be expressed in σ
max
.
available elastic energy per surfaceunity [N m
−1
]
U
i
=
1
S
_
x=λ/2
x=0
f(x) dx
16
=
_
x=λ/2
x=0
σ
max
sin
_
2πx
λ
_
dx
= σ
max
λ
π
[Nm
−1
]
required surface energy
U
a
= 2γ [Nm
−1
]
energy balance at fracture
U
i
= U
a
→ λ =
2πγ
σ
max
→ σ = σ
max
sin
_
x
γ
σ
max
_
The argument of the sine function is assumed to be small enough to allow a linear approxima
tion. The displacement x between the atomic planes is expressed in the linear strain ε. The
macroscopic Young’s modulus E of the material is introduced as the derivative of the stress
w.r.t. the linear strain in the undeformed state. This leads to a relation for the theoretical
strength σ
th
= σ
max
.
linearization
σ = σ
max
sin
_
x
γ
σ
max
_
≈
x
γ
σ
2
max
linear strain of atomic bond
ε =
x
a
0
→ x = εa
0
→ σ =
εa
0
γ
σ
2
max
elastic modulus
E =
_
dσ
dε
_¸
¸
¸
¸
x=0
=
_
dσ
dx
a
0
_¸
¸
¸
¸
x=0
= σ
2
max
a
0
γ
→
σ
max
=
_
Eγ
a
0
theoretical strength
σ
th
=
_
Eγ
a
0
2.3.1 Discrepancy with experimental observations
The surface energy γ does not diﬀer much for various solid materials and approximately equals
1 Jm
−2
. (Diamond is an exception with γ = 5 Jm
−2
.) The equilibrium distance a
0
between
atoms, is also almost the same for solids (about 10
−10
m).
The table below lists values of theoretical strength and experimental fracture stress for
some materials. It is clear that there is a large discrepancy between the two values : the
theoretical strength is much too high. The reason for this deviation has been discovered by
Griﬃth in 1921.
17
a
0
[m] E [GPa] σ
th
[GPa] σ
b
[MPa] σ
th
/σ
b
glass 3 ∗ 10
−10
60 14 170 82
steel 10
−10
210 45 250 180
silica ﬁbers 10
−10
100 31 25000 1.3
iron whiskers 10
−10
295 54 13000 4.2
silicon whiskers 10
−10
165 41 6500 6.3
alumina whiskers 10
−10
495 70 15000 4.7
ausformed steel 10
−10
200 45 3000 15
piano wire 10
−10
200 45 2750 16.4
σ
th
≫σ
b
2.3.2 Griﬃth’s experiments
In 1921 Griﬃth determined experimentally the fracture stress σ
b
of glass ﬁbers as a function
of their diameter. For d > 20 µm the bulk strength of 170 MPa was found. However, σ
b
approached the theoretical strength of 14000 MPa in the limit of zero thickness.
Griﬃth new of the earlier (1913) work of Inglis [32], who calculated stress concentrations
at circular holes in plates, being much higher than the nominal stress. He concluded that in
his glass ﬁbers such stress concentrations probably occurred around defects and caused the
discrepancy between theoretical and experimental fracture stress. He reasoned that for glass
ﬁbers with smaller diameters, there was less volume and less chance for a defect to exist in the
specimen. In the limit of zero volume there would be no defect and the theoretical strength
would be found experimentally. Griﬃth published his work in 1921 and his paper [28] can be
seen as the birth of Fracture Mechanics. It was shown in 1976 by Parratt, March en Gordon,
that surface defects instead of volume defects were the cause for the limiting strength.
The ingenious insight that strength was highly inﬂuenced by defects has lead to the shift
of attention to the behavior of cracks and the formulation of crack growth criteria. Fracture
Mechanics was born!
σ
b
11000
170
10 20
d [µ]
[MPa]
Fig. 2.9 : Fracture strength of glass ﬁbers in relation to their thickness.
18
2.3.3 Crack loading modes
Irwin was one of the ﬁrst to study the behavior of cracks. He introduced three diﬀerent
loading modes, which are still used today [33].
Mode I Mode II Mode III
Fig. 2.10 : Three standard loading modes of a crack.
Mode I = opening mode
Mode II = sliding mode
Mode III = tearing mode
Chapter 3
Experimental techniques
To predict the behavior of a crack, it is essential to know its location, geometry and dimen
sions. Experiments have to be done to reveal these data.
Experimental techniques have been and are still being developed. Some of these proce
dures use physical phenomena to gather information about a crack. Other techniques strive
towards visualization of the crack.
3.1 Surface cracks
One of the most simple techniques to reveal surface cracks is based on dye penetration into
the crack due to capillary ﬂow of the dye. Although it can be applied easily and onsite, only
surface cracks can be detected.
Other simple procedures are based on the observation of the disturbance of the magnetic
or electric ﬁeld, caused by a crack. Magnetic ﬁelds can be visualized with magnetic particles
and electric ﬁelds by the use of inertances. Only cracks at or just below the surface can be
detected in this way.
Fig. 3.1 : Dye penetration in the stiﬀening cone of a turbine.
(Source: Internet site www.ventioelde.de (2009))
19
20
3.2 Electrical resistance
A crack is a discontinuity in the material and as such diminishes the crosssectional area.
This may be associated with an increase of the electrical resistance, which can be measured
for metallic materials and carbon composites.
3.3 Xray
Direct visualization of a crack can be done using electromagnetic waves. Xrays are routinely
used to control welds.
Fig. 3.2 : Xray robots inside and outside of a pipe searching for cracks.
(Source: Internet site University of Strathclyde (2006))
3.4 Ultrasound
Visualization is also possible with sound waves. This is based on the measurement of the
distance over which a wave propagates from its source via the reﬂecting crack surface to a
detector.
pi¨ezoel. crystal
sensor
wave
∆t
S in
out
t
Fig. 3.3 : Ultrasound crack detection.
21
3.5 Acoustic emission
The release of energy in the material due to crack generation and propagation, results in sound
waves (elastic stress waves), which can be detected at the surface. There is a correlation of
their amplitude and frequency with failure phenomena inside the material. This acoustic
emission (AE) is much used in laboratory experiments.
3.6 Adhesion tests
Many experimental techniques are used to determine adhesion strength of surface layers on
a substrate. Some of them are illustrated below.
blade wedge test
peel test (0
o
and 90
o
)
bending test
scratch test
indentation test laser blister test
pressure blister test
fatigue friction test
22
Chapter 4
Fracture energy
4.1 Energy balance
Abandoning the inﬂuence of thermal eﬀects, the ﬁrst law of thermodynamics can be formu
lated for a unit material volume :
the total amount of mechanical energy that is supplied to a material volume per unit of
time (
˙
U
e
) must be transferred into internal energy (
˙
U
i
), surface energy (
˙
U
a
), dissipated
energy (
˙
U
d
) and kinetic energy (
˙
U
k
).
The internal energy is the elastically stored energy. The surface energy changes, when new
free surface is generated, e.g. when a crack propagates. The kinetic energy is the result of
material velocity. The dissipation may have various characteristics, but is mostly due to
friction and plastic deformation. It results in temperature changes.
Instead of taking time derivatives (˙) to indicate changes, we can also use another state
variable, e.g. the crack surface area A. Assuming the crack to be through the thickness of
a plate, whose thickness B is uniform and constant, we can also use the crack length a as a
state variable.
The internally stored elastic energy is considered to be an energy source and therefore
it is moved to the lefthand side of the energy balance equation.
a
A = Ba
B = thickness
Fig. 4.1 : Plate with a line crack of length a.
23
24
˙
U
e
=
˙
U
i
+
˙
U
a
+
˙
U
d
+
˙
U
k
[Js
−1
]
d
dt
( ) =
dA
dt
d
dA
( ) =
˙
A
d
dA
( ) = ˙ a
d
da
( )
dU
e
da
=
dU
i
da
+
dU
a
da
+
dU
d
da
+
dU
k
da
[Jm
−1
]
dU
e
da
−
dU
i
da
=
dU
a
da
+
dU
d
da
+
dU
k
da
[Jm
−1
]
4.2 Griﬃth’s energy balance
It is now assumed that the available external and internal energy is transferred into surface
energy. Dissipation and kinetic energy are neglected. This results in the socalled Griﬃth
energy balance.
After division by the plate thickness B, the lefthand side of the equation is called the
energy release rate G and the righthand side the crack resistance force R, which equals 2γ,
where γ is the surface energy of the material.
energy balance
dU
e
da
−
dU
i
da
=
dU
a
da
energy release rate G =
1
B
_
dU
e
da
−
dU
i
da
_
[Jm
−2
]
crack resistance force R =
1
B
_
dU
a
da
_
= 2γ [Jm
−2
]
Griﬃth’s crack criterion G = R = 2γ [Jm
−2
]
According to Griﬃth’s energy balance, a crack will grow, when the energy release rate equals
the crack resistance force.
This is illustrated with the following example, where a plate is loaded in tension and
ﬁxed at its edges. An edge crack of length a is introduced in the plate. Because the edges are
clamped, the reaction force, due to prestraining, does not do any work, so in this case of ﬁxed
grips we have dU
e
= 0. The material volume, where elastic energy is released, is indicated as
the shaded area around the crack. The released elastic energy is a quadratic function of the
crack length a. The surface energy, which is associated with the crack surface is 2γa.
Both surface energy U
a
and internal energy U
i
are plotted as a function of the crack
length. The initial crack length a is indicated in the ﬁgure and it must be concluded that for
crack growth da, the increment in surface energy is higher than the available (decrease of)
internal energy. This crack therefore cannot growth and is stable.
When we gradually increase the crack length, the crack becomes unstable at the length
a
c
. In that case the crack growth criterion is just met (G = R). The increase of surface
energy equals the decrease of the internal energy.
25
a
a
c
a
c
da
2γ
G, R
U
i
U
a
needed
available
dU
e
= 0
Fig. 4.2 : Illustration of Griﬃth’s energy balance criterion.
−
dU
i
da
<
dU
a
da
→ no crack growth
−
dU
i
da
>
dU
a
da
→ unstable crack growth
−
dU
i
da
=
dU
a
da
→ critical crack length
4.3 Griﬃth stress
We consider a crack of length 2a in an ”inﬁnite” plate with uniform thickness B [m]. The
crack is loaded in mode I by a nominal stress σ [Nm
−2
], which is applied on edges at large
distances from the crack.
Because the edges with the applied stress are at a far distance from the crack, their
displacement will be very small when the crack length changes slightly. Therefore it is assumed
that dU
e
= 0 during crack propagation. It is further assumed that the elastic energy in the
elliptical area is released, when the crack with length a is introduced. The energy release rate
and the crack resistance force can now be calculated. According to Griﬃth’s energy balance,
the applied stress σ and the crack length a are related. From this relation we can calculate
the Griﬃth stress σ
gr
and also the critical crack length a
c
.
26
2a
a
x
σ
σ
y
thickness B
Fig. 4.3 : Elliptical region, which is unloaded due to the central crack of length 2a.
U
i
= 2πa
2
B
1
2
σ
2
E
; U
a
= 4aB γ [Nm = J]
Griﬃth’s energy balance (dU
e
= 0)
G = −
1
B
_
dU
i
da
_
=
1
B
_
dU
a
da
_
= R → 2πa
σ
2
E
= 4γ [Jm
−2
]
Griﬃth stress σ
gr
=
_
2γE
πa
critical crack length a
c
=
2γE
πσ
2
4.3.1 Discrepancy with experimental observations
When the Griﬃth stress is compared to the experimental critical stress for which a crack of
length a will propagate, it appears that the Griﬃth stress is much too small: it underestimates
the strength.
The reason for this discrepancy is mainly due to the fact that in the Griﬃth energy
balance the dissipation is neglected. This can be concluded from the comparison of the crack
resistance force R = 4γ and the measured critical energy release rate G
c
. For materials which
are very brittle (e.g. glass) and thus show little dissipation during crack growth, the diﬀerence
between R and G
c
is not very large. For ductile materials, showing much dissipation (e.g.
metal alloys), G
c
can be 10
5
times R.
Griﬃth’s energy balance can be used in practice, when the calculated energy release rate
is compared to a measured critical value G
c
.
energy balance G =
1
B
_
dU
e
da
−
dU
i
da
_
= R = G
c
27
critical crack length a
c
=
G
c
E
2πσ
2
Griﬃth’s crack criterion G = G
c
4.4 Compliance change
The energy release rate can be calculated from the change in stiﬀness due to the elongation
of a crack. It is common practice to use the compliance C instead of the stiﬀness and the
relation between G and the change of C will be derived for a plate with an edge crack, which
is loaded by a force F in point P. When the crack length increases from a to a + da, two
extreme situations can be considered :
• ﬁxed grips : point P is not allowed to move (u = 0),
• constant load : force F is kept constant.
P
F
u
F
u
a +da a
P
Fig. 4.4 : Edge crack in a plate loaded by a force F.
a a
a +da a +da
F F
u u
dU
i
dU
i
dU
e
Fig. 4.5 : Internal and external energy for ﬁxed grips (left) and constant load (right).
28
4.4.1 Fixed grips
Using the ﬁxed grips approach, it is obvious that dU
e
= 0. The force F will decrease (dF < 0)
upon crack growth and the change of the internally stored elastic energy can be expressed in
the displacement u and the change dF. The energy release rate can be calculated according
to its deﬁnition.
ﬁxed grips : dU
e
= 0
dU
i
= U
i
(a +da) −U
i
(a) (< 0)
=
1
2
(F +dF)u −
1
2
Fu
=
1
2
udF
Griﬃth’s energy balance
G = −
1
2B
u
dF
da
=
1
2B
u
2
C
2
dC
da
=
1
2B
F
2
dC
da
4.4.2 Constant load
With the constant load approach, the load will supply external work, when the crack propa
gates and the point P moves over a distance du. Also the elastic energy will diminish and this
change can be expressed in F and du. The energy release rate can be calculated according to
its deﬁnition.
constant load
dU
e
= U
e
(a +da) −U
e
(a) = Fdu
dU
i
= U
i
(a +da) −U
i
(a) (> 0)
=
1
2
F(u +du) −
1
2
Fu
=
1
2
Fdu
Griﬃth’s energy balance
G =
1
2B
F
du
da
=
1
2B
F
2
dC
da
4.4.3 Experiment
It can be concluded that the energy release rate can be calculated from the change in com
pliance and that the result for the ﬁxed grip approach is exactly the same as that for the
constant load method.
In reality the loading of the plate may not be purely according to the extreme cases, as
is shown in the ﬁgure. From such a real experiment the energy release rate can be determined
from the forcedisplacement curve.
29
a
1
u
F
a
2
a
P
u
F
a
4
a
3
Fig. 4.6 : Experimental forcedisplacement curve during crack growth.
G =
shaded area
a
4
−a
3
1
B
4.4.4 Examples
As an example the energy release rate is calculated for the socalled double cantilever beam,
shown in the ﬁgure.
Using linear elastic beam theory, the opening ∆u of the crack can be expressed in the
force F, the crack length a, beam thickness B, beam height 2h and Young’s modulus E. The
compliance is the ratio of opening and force. Diﬀerentiating the compliance w.r.t. the crack
length results in the energy release rate G, which appears to be quadratic in a.
a
F
2h
u
B
F
u
Fig. 4.7 : Beam with a central crack loaded in Mode I.
30
u =
Fa
3
3EI
=
4Fa
3
EBh
3
C =
∆u
F
=
2u
F
=
8a
3
EBh
3
→
dC
da
=
24a
2
EBh
3
→
G =
1
B
_
1
2
F
2
dC
da
_
=
12F
2
a
2
EB
2
h
3
[J m
−2
]
G
c
= 2γ → F
c
=
B
a
_
1
6
γEh
3
The question arises for which beam geometry, the energy release rate will be constant, so no
function of the crack length. It is assumed that the height of the beam is an exponential
function of a. Applying again formulas from linear elastic beam theory, it can be derived for
which shape G is no function of a.
a
h
Fig. 4.8 : Cracked beam with variable crosssection.
C =
∆u
F
=
2u
F
=
8a
3
EBh
3
→
dC
da
=
24a
2
EBh
3
choice : h = h
0
a
n
→
u =
Fa
3
3(1 −n)EI
=
4Fa
3
(1 −n)EBh
3
=
4Fa
3(1−n)
(1 −n)EBh
3
0
C =
2u
F
=
8a
3(1−n)
(1 −n)EBh
3
0
→
dC
da
=
24a
(2−3n)
EBh
3
0
dC
da
constant for n =
2
3
→ h = h
0
a
2
3
Chapter 5
Stress concentrations
In this chapter attention is given to stress concentrations due to holes. Throughout this
chapter it is assumed that the material behavior is linear elastic and isotropic. The stress
state is calculated using Airy stress functions. The ﬁrst sections present a summary of the
theory of linear elasticity.
5.1 Deformation and strain
The basic problem in mechanics is the prediction of the mechanical behavior of a body when
it is subjected to an external load. We want to predict the deformation, the deformation
rate and the stresses in the material. (Temperature is not considered here, but could also be
determined, making the behavior thermomechanical.)
To set up a mathematical model, which can be used to solve the above problem, we
must be able to identify the position of material points. For this purpose we use position
vectors :
X in the undeformed state and x in the deformed state. The displacement during
deformation is obviously the diﬀerence between x and
X. In the mathematics of the next
sections, we use the components of the vectorial (and tensorial) variables w.r.t. a coordinate
system. These components are indicated by indices, which take values in the rage {1, 2, 3}
for threedimensional deformations.
In continuum mechanics we assume that the body is perfect, i.e. that it does not contain
voids, cracks or other imperfections. This implies that relevant variables, like the displace
ment, are a continuous function of the location within the body. In this way the diﬀerence
vector between two adjacent points in undeformed and deformed state can be related.
Q
P
Q
P
x
x +dx
X +d
X
X
u
e
3
e
2
e
1
Fig. 5.1 : Deformation of a continuum.
31
32
x
i
= X
i
+u
i
(X
i
)
x
i
+dx
i
= X
i
+dX
i
+u
i
(X
i
+dX
i
) = X
i
+dX
i
+u
i
(X
i
) +u
i,j
dX
j
dx
i
= dX
i
+u
i,j
dX
j
= (δ
ij
+u
i,j
)dX
j
The local deformation, i.e. the deformation of a small material volume in a material point, is
described by strains. When elongations and rotations are large, the GreenLagrange strains
γ
ij
are commonly used. For small elongations and rotations, these can be linearized to the
linear strains ε
ij
.
ds
2
= dx
i
dx
i
= [(δ
ij
+u
i,j
)dX
j
][(δ
ik
+u
i,k
)dX
k
]
= (δ
ij
δ
ik
+δ
ij
u
i,k
+u
i,j
δ
ik
+u
i,j
u
i,k
)dX
j
dX
k
= (δ
jk
+u
j,k
+u
k,j
+u
i,j
u
i,k
)dX
j
dX
k
= (δ
ij
+u
i,j
+u
j,i
+u
k,i
u
k,j
)dX
i
dX
j
= dX
i
dX
i
+ (u
i,j
+u
j,i
+u
k,i
u
k,j
)dX
i
dX
j
= dS
2
+ (u
i,j
+u
j,i
+u
k,i
u
k,j
)dX
i
dX
j
ds
2
−dS
2
= (u
i,j
+u
j,i
+u
k,i
u
k,j
)dX
i
dX
j
= 2γ
ij
dX
i
dX
j
GreenLagrange strains γ
ij
=
1
2
(u
i,j
+u
j,i
+u
k,i
u
k,j
)
linear strains ε
ij
=
1
2
(u
i,j
+u
j,i
)
Because three displacement components are the basis for the deﬁnition of six strain compo
nents, the latter cannot be independent. The compatibility relations express this dependency.
2ε
12,12
−ε
11,22
−ε
22,11
= 0
2ε
23,23
−ε
22,33
−ε
33,22
= 0
2ε
31,31
−ε
33,11
−ε
11,33
= 0
ε
11,23
+ε
23,11
−ε
31,12
−ε
12,13
= 0
ε
22,31
+ε
31,22
−ε
12,23
−ε
23,21
= 0
ε
33,12
+ε
12,33
−ε
23,31
−ε
31,32
= 0
33
5.2 Stress
When deformation is not completely unconstrained, it provokes stresses in the material. The
stress vector p on a plane in a material point, can be calculated from the unity normal
vector n on the plane and the Cauchy stress tensor. Using index notation in a coordinate
system, the components of p can be related to those of n by the Cauchy stress components
σ
ij
(i, j = 1, 2, 3). These stress components can be represented as the stress vectors on
the sides of an inﬁnitesimal stress cube in the material point, with its sides normal to the
coordinate axes.
unity normal vector n = n
i
e
i
stress vector p = p
i
e
i
Cauchy stress components p
i
= σ
ij
n
j
σ
13
σ
33
σ
22
2
3
σ
11
1
σ
21
σ
31
σ
32
σ
12
σ
23
Fig. 5.2 : Stress components on a stress cube.
5.3 Linear elastic material behavior
For very small deformations, the mechanical behavior of a solid material is reversible and
characterized by a linear relation between stress and strain components. The relation is given
by the stiﬀness parameters C
ijkl
(i, j, k, l = 1, 2, 3) and is referred to as Hooke’s law.
Due to symmetry of stress and strain components, there are no more than 21 independent
stiﬀness parameters. Due to material symmetry – crystal structure – the elastic behavior of
most materials can be described by considerably less material parameters. For isotropic
materials, two material parameters are suﬃcient : Young’s modulus E and Poisson’s ratio ν.
σ
ij
= C
ijkl
ε
lk
34
Hooke’s law for isotropic materials
For isotropic materials, Hooke’s law can be written in index notation, relating six strain
components to six stress components and vice versa. When these components are stored in
columns, Hooke’s law is represented by a sixbysix matrix.
σ
ij
=
E
1 +ν
_
ε
ij
+
ν
1 −2ν
δ
ij
ε
kk
_
i = 1, 2, 3
ε
ij
=
1 +ν
E
_
σ
ij
−
ν
1 +ν
δ
ij
σ
kk
_
i = 1, 2, 3
_
¸
¸
¸
_
σ
11
σ
22
σ
33
σ
12
σ
23
σ
31
_
¸
¸
¸
_
= α
_
¸
¸
¸
_
1 −ν ν ν 0 0 0
ν 1 −ν ν 0 0 0
ν ν 1 −ν 0 0 0
0 0 0 1 −2ν 0 0
0 0 0 0 1 −2ν 0
0 0 0 0 0 1 −2ν
_
¸
¸
¸
_
_
¸
¸
¸
_
ε
11
ε
22
ε
33
ε
12
ε
23
ε
31
_
¸
¸
¸
_
α = E/[(1 +ν)(1 −2ν)]
_
¸
¸
¸
¸
¸
¸
_
ε
11
ε
22
ε
33
ε
12
ε
23
ε
31
_
¸
¸
¸
¸
¸
¸
_
=
1
E
_
¸
¸
¸
¸
¸
¸
_
1 −ν −ν 0 0 0
−ν 1 −ν 0 0 0
−ν −ν 1 0 0 0
0 0 0 1 +ν 0 0
0 0 0 0 1 +ν 0
0 0 0 0 0 1 +ν
_
¸
¸
¸
¸
¸
¸
_
_
¸
¸
¸
¸
¸
¸
_
σ
11
σ
22
σ
33
σ
12
σ
23
σ
31
_
¸
¸
¸
¸
¸
¸
_
5.4 Equilibrium equations
A material, which is subjected to external loads, will deform and the stresses in the deformed
state have to satisfy the equilibrium equations in each point. The three equilibrium equations
for forces are partial diﬀerential equations, which can only be solved with proper boundary
conditions. The three equilibrium equations for moments indicate that the Cauchy stress
components are symmetric.
We have assumed here that the acceleration of the material points can be neglected.
35
σ
13
+σ
13,3
dx
3
σ
33
+σ
33,3
dx
3
σ
23
σ
13
σ
33
3
2
1
σ
23
+σ
23,3
dx
3
σ
31
+σ
31,1
dx
1
σ
22
σ
21
σ
22
+σ
22,2
dx
2
σ
12
σ
32
σ
11
σ
31
σ
12
+σ
12,2
dx
2
σ
32
+σ
32,2
dx
2
σ
11
+σ
11,1
dx
1
σ
21
+σ
21,1
dx
1
Fig. 5.3 : Stress components on a stress cube with a volume load.
volume load ρq
i
force equilibrium σ
ij,j
+ρq
i
= 0 i = 1, 2, 3
moment equilibrium σ
ij
= σ
ji
5.5 Plane stress
When stresses are zero on a certain plane, which we take to be the {1, 2}plane in this case,
we have a plane stress state.
Only the three independent stress components in the plane have to be considered. They
have to satisfy two equilibrium equations. The three strain components in the plane are
related by one compatibility relation. Hooke’s law for linear elastic material behavior relates
the three stress components to the three strain components, with a 3 × 3 stiﬀness matrix or
compliance matrix.
equilibrium (q
i
= 0) σ
11,1
+σ
12,2
= 0 ; σ
21,1
+σ
22,2
= 0
compatibility 2ε
12,12
−ε
11,22
−ε
22,11
= 0
Hooke’s law
σ
ij
=
E
1 +ν
_
ε
ij
+
ν
1 −ν
δ
ij
ε
kk
_
; ε
ij
=
1 +ν
E
_
σ
ij
−
ν
1 +ν
δ
ij
σ
kk
_
i = 1, 2
Hooke’s law in matrix notation
36
_
_
ε
11
ε
22
ε
12
_
_
=
1
E
_
_
1 −ν 0
−ν 1 0
0 0 1 +ν
_
_
_
_
σ
11
σ
22
σ
12
_
_
_
_
σ
11
σ
22
σ
12
_
_
=
E
1 −ν
2
_
_
1 ν 0
ν 1 0
0 0 1 −ν
_
_
_
_
ε
11
ε
22
ε
12
_
_
ε
33
= −
ν
E
(σ
11
+σ
22
) = −
ν
1 −ν
(ε
11
+ε
22
)
ε
13
= ε
23
= 0
5.6 Plane strain
In the plane strain situation, the deformation of the material is such that there is no elongation
and shear in and perpendicular to one direction, which we here take to be the 3direction,
perpendicular to the {1, 2}plane.
Only three strain components remain, which have to satisfy one compatibility relation.
Three stress components have to be solved from two equilibrium equations. Hooke’s law can
be simpliﬁed and represented with 3 ×3 stiﬀness and compliance matrices.
equilibrium (q
i
= 0) σ
11,1
+σ
12,2
= 0 ; σ
21,1
+σ
22,2
= 0
compatibility 2ε
12,12
−ε
11,22
−ε
22,11
= 0
Hooke’s law
ε
ij
=
1 +ν
E
(σ
ij
−νδ
ij
σ
kk
) ; σ
ij
=
E
1 +ν
_
ε
ij
+
ν
1 −2ν
δ
ij
ε
kk
_
i = 1, 2
Hooke’s law in matrix notation
_
_
σ
11
σ
22
σ
12
_
_
=
E
(1 +ν)(1 −2ν)
_
_
1 −ν ν 0
ν 1 −ν 0
0 0 1 −2ν
_
_
_
_
ε
11
ε
22
ε
12
_
_
_
_
ε
11
ε
22
ε
12
_
_
=
1 +ν
E
_
_
1 −ν −ν 0
−ν 1 −ν 0
0 0 1
_
_
_
_
σ
11
σ
22
σ
12
_
_
σ
33
=
Eν
(1 +ν)(1 −2ν)
(ε
11
+ε
22
) = ν (σ
11
+σ
22
)
σ
13
= σ
23
= 0
37
5.7 Displacement method
The displacement method is based on transforming the equilibrium equations to diﬀerential
equations for the displacement components.
First, the stressstrain relation according to Hooke’s law are substituted into the equi
librium equations, leading to two diﬀerential equations for the strain components. Next, the
straindisplacement relations are substituted, resulting in the ﬁnal two diﬀerential equations
for the two displacement components. These equations can be solved with proper boundary
conditions (BC’s). Analytical solutions can only be determined for simple cases.
σ
ij,j
= 0
σ
ij
=
E
1 +ν
_
ε
ij
+
ν
1 −2ν
δ
ij
ε
kk,j
_
_
¸
¸
_
¸
¸
_
E
1 +ν
_
ε
ij,j
+
ν
1 −2ν
δ
ij
ε
kk,j
_
= 0
ε
ij
=
1
2
(u
i,j
+u
j,i
)
_
¸
¸
_
¸
¸
_
E
1 +ν
1
2
(u
i,jj
+u
j,ij
) +
Eν
(1 +ν)(1 −2ν)
δ
ij
u
k,kj
= 0
BC’s
_
¸
¸
_
¸
¸
_
u
i
→ ε
ij
→ σ
ij
5.8 Stress function method
According to the displacement method, the equilibrium equation is transformed to a diﬀer
ential equation in the displacements. In the socalled stress function method, the equilibrium
is transformed to a diﬀerential equation in the Airy stress function ψ.
The stress components are related to derivatives of the stress function, such that they
satisfy the equilibrium equations. Using Hooke’s law, the strain components are expressed in
derivatives of ψ. The strain components are then substituted in the compatibility equations.
Here only twodimensional problems are considered, so only one compatibility equation is
relevant.
The result is a fourthorder diﬀerential equation for the stress function ψ, which is called
the biharmonic equation. It can be written in short form using the Laplace operator ∇
2
.
With proper boundary conditions (4), the biharmonic equation can be solved (theoret
ically). When the stress function is known, the stresses, strains and displacements can be
derived rather straightforwardly.
ψ(x
1
, x
2
) → σ
ij
= −ψ
,ij
+δ
ij
ψ
,kk
→ σ
ij,j
= 0
ε
ij
=
1 +ν
E
(σ
ij
−νδ
ij
σ
kk
)
_
¸
_
¸
_
38
ε
ij
=
1 +ν
E
{−ψ
,ij
+ (1 −ν)δ
ij
ψ
,kk
}
2ε
12,12
−ε
11,22
−ε
22,11
= 0
_
¸
_
¸
_
2ψ
,1122
+ψ
,2222
+ψ
,1111
= 0 →
(ψ
,11
+ψ
,22
)
,11
+ (ψ
,11
+ψ
,22
)
,22
= 0
Laplace operator : ∇
2
=
∂
2
∂x
2
1
+
∂
2
∂x
2
2
= ( )
11
+ ( )
22
_
¸
¸
_
¸
¸
_
→
biharmonic equation ∇
2
(∇
2
ψ) = ∇
4
ψ = 0
BC’s
_
ψ → σ
ij
→ ε
ij
→ u
i
Cylindrical coordinates
The biharmonic equation has been derived and expressed in Cartesian coordinates. To get
an expression in cylindrical coordinates, the base vectors of the cylindrical coordinate system
are expressed in the base vectors of the Cartesian coordinate system. It is immediately clear
that cylindrical base vectors are a function of the cylindrical coordinate θ.
e
3
θ
e
z
e
t
e
r
x
y
z
r
e
1
e
2
Fig. 5.4 : Cartesian and cylindrical coordinate system.
vector bases {e
1
, e
2
, e
3
} → {e
r
, e
t
, e
z
}
e
r
= e
r
(θ) = e
1
cos θ +e
2
sinθ
e
t
= e
t
(θ) = −e
1
sinθ +e
2
cos θ
39
∂
∂θ
{e
r
(θ)} = e
t
(θ) ;
∂
∂θ
{e
t
(θ)} = −e
r
(θ)
Laplace operator
The Laplace operator is a scalar operator and can be written as the inner product of the
gradient operator, which is a vector operator.
gradient operator
∇ = e
r
∂
∂r
+e
t
1
r
∂
∂θ
+e
z
∂
∂z
Laplace operator ∇
2
=
∇·
∇ =
∂
2
∂r
2
+
1
r
∂
∂r
+
1
r
2
∂
2
∂θ
2
+
∂
2
∂z
2
twodimensional ∇
2
=
∂
2
∂r
2
+
1
r
∂
∂r
+
1
r
2
∂
2
∂θ
2
Biharmonic equation
The biharmonic equation can now be written in cylindrical coordinates. Also the stress com
ponents, which are derivatives of the stress function, can be written in cylindrical coordinates.
The latter expression can best be derived from the expression for the stress tensor and using
the expression for the gradient operator in cylindrical coordinates.
biharmonic equation
_
∂
2
∂r
2
+
1
r
∂
∂r
+
1
r
2
∂
2
∂θ
2
__
∂
2
ψ
∂r
2
+
1
r
∂ψ
∂r
+
1
r
2
∂
2
ψ
∂θ
2
_
= 0
stress components
σ
rr
=
1
r
∂ψ
∂r
+
1
r
2
∂
2
ψ
∂θ
2
σ
tt
=
∂
2
ψ
∂r
2
σ
rt
=
1
r
2
∂ψ
∂θ
−
1
r
∂
2
ψ
∂r∂θ
= −
∂
∂r
_
1
r
∂ψ
∂θ
_
5.9 Circular hole in ’inﬁnite’ plate
A circular hole with diameter 2a is located in a plate with uniform thickness. Dimensions
of the plate are much larger than 2a. The plate is loaded with a uniform stress σ in the
global xdirection. The center of the hole is the origin of a cylindrical coordinate system. To
determine the stress components as a function of r and θ, the biharmonic equation must be
solved to determine the Airy stress function ψ(r, θ) for this problem.
40
2a
x
y
r
σ σ
θ
Fig. 5.5 : A circular hole in a plate, loaded with a uniform stress σ in xdirection.
Load transformation
At a circular crosssection at a distance r = b ≫a, the uniform stress σ can be decomposed
in a radial stress σ
rr
and a shear stress σ
rt
, which can be related to σ using equilibrium.
The total load at the crosssection r = b is the sum of a radial load and a load that
depends on θ. For both loadcases, the stress ﬁeld for b ≤ r ≤ a will be determined, after
which the combined solution follows from superposition.
σ
rr
σ
rt
σ
σ
σ
rt
σ
rr
θ
2b
2a
Fig. 5.6 : Transformation of the xload in radial and circumferential loads.
σ
rr
(r = b, θ) =
1
2
σ +
1
2
σ cos(2θ)
σ
rt
(r = b, θ) = −
1
2
σ sin(2θ)
41
two load cases
I. σ
rr
(r = a) = σ
rt
(r = a) = 0
σ
rr
(r = b) =
1
2
σ ; σ
rt
(r = b) = 0
II. σ
rr
(r = a) = σ
rt
(r = a) = 0
σ
rr
(r = b) =
1
2
σ cos(2θ) ; σ
rt
(r = b) = −
1
2
σ sin(2θ)
Load case I
In the ﬁrst loadcase, the boundary at r = b is subjected to a radial load σ
rr
=
1
2
σ. Because
σ
rr
is not depending on θ, the problem is axisymmetric and ψ is only a function of r. The
stress components can be expressed in this function. The biharmonic equation is an ordinary
diﬀerential equation in the variable r.
Airy function ψ = f(r)
stress components
σ
rr
=
1
r
∂ψ
∂r
+
1
r
2
∂
2
ψ
∂θ
2
=
1
r
df
dr
σ
tt
=
∂
2
ψ
∂r
2
=
d
2
f
dr
2
σ
rt
= −
∂
∂r
_
1
r
∂ψ
∂θ
_
= 0
biharmonic equation
_
d
2
dr
2
+
1
r
d
dr
__
d
2
f
dr
2
+
1
r
df
dr
_
= 0
A general solution for the diﬀerential equation has four integration constants. These have
to be determined from the boundary conditions for this loadcase. However, the boundary
conditions for σ
rt
cannot be used as this stress component is zero everywhere. Instead the
compatibility equation is used, which relates the strain components ε
rr
and ε
tt
. Also b ≫ a
is used in the calculation.
general solution ψ(r) = Aln r +Br
2
ln r +Cr
2
+D
stresses σ
rr
=
A
r
2
+B(1 + 2 ln r) + 2C
σ
tt
= −
A
r
2
+B(3 + 2 ln r) + 2C
strains (from Hooke’s law for plane stress)
ε
rr
=
1
E
_
A
r
2
(1 +ν) +B{(1 −3ν) + 2(1 −ν) ln r} + 2C(1 −ν)
_
ε
tt
=
1
E
1
r
_
−
A
r
(1 +ν) +B{(3 −ν)r + 2(1 −ν)r ln r} + 2C(1 −ν)r
_
42
compatibility ε
rr
=
du
dr
=
d(r ε
tt
)
dr
→ B = 0
2 BC’s and b ≫a → A and C →
σ
rr
=
1
2
σ(1 −
a
2
r
2
) ; σ
tt
=
1
2
σ(1 +
a
2
r
2
) ; σ
rt
= 0
Load case II
The second loadcase is again deﬁned by the stress components at r = a and r = b. At the
boundary r = b the stresses are harmonic functions of the coordinate θ. Confronting the
general expressions for stresses σ
rr
and σ
rt
with these prescribed values at r = b, indicates
that the Airy function ψ(r, θ) must be written as the product of a function g(r) and cos(2θ).
(The term cos(2θ) remains after two subsequent diﬀerentiations w.r.t. θ.) Substitution in the
biharmonic equation results in a diﬀerential equation for g(r) and expressions for the stress
components.
Airy function ψ(r, θ) = g(r) cos(2θ)
stress components
σ
rr
=
1
r
∂ψ
∂r
+
1
r
2
∂
2
ψ
∂θ
2
; σ
tt
=
∂
2
ψ
∂r
2
σ
rt
=
1
r
2
∂ψ
∂θ
−
1
r
∂
2
ψ
∂r∂θ
= −
∂
∂r
_
1
r
∂ψ
∂θ
_
biharmonic equation
_
∂
2
∂r
2
+
1
r
∂
∂r
+
1
r
2
∂
2
∂θ
2
___
d
2
g
dr
2
+
1
r
dg
dr
−
4
r
2
g
_
cos(2θ)
_
= 0 →
_
d
2
dr
2
+
1
r
d
dr
−
4
r
2
__
d
2
g
dr
2
+
1
r
dg
dr
−
4
r
2
g
_
cos(2θ) = 0
A general solution for the diﬀerential equation has four integration constants. These have
to be determined from the boundary conditions for this loadcase. Using b ≫ a, the stress
components σ
rr
, σ
tt
and σ
rt
can be determined as a function of r and θ.
general solution g = Ar
2
+Br
4
+C
1
r
2
+D →
ψ =
_
Ar
2
+Br
4
+C
1
r
2
+D
_
cos(2θ)
stresses σ
rr
= −
_
2A +
6C
r
4
+
4D
r
2
_
cos(2θ)
σ
tt
=
_
2A + 12Br
2
+
6C
r
4
_
cos(2θ)
43
σ
rt
=
_
2A + 6Br
2
−
6C
r
4
−
2D
r
2
_
sin(2θ)
4 BC’s and b ≫a → A,B,C and D →
σ
rr
=
1
2
σ
_
1 +
3a
4
r
4
−
4a
2
r
2
_
cos(2θ)
σ
tt
= −
1
2
σ
_
1 +
3a
4
r
4
_
cos(2θ)
σ
rt
= −
1
2
σ
_
1 −
3a
4
r
4
+
2a
2
r
2
_
sin(2θ)
Stresses for total load
The stress components σ
rr
, σ
tt
and σ
rt
resulting from the total load at r = b ≫ a, are the
sum of the solutions for the separate loadcases.
σ
rr
=
σ
2
__
1 −
a
2
r
2
_
+
_
1 +
3a
4
r
4
−
4a
2
r
2
_
cos(2θ)
_
σ
tt
=
σ
2
__
1 +
a
2
r
2
_
−
_
1 +
3a
4
r
4
_
cos(2θ)
_
σ
rt
= −
σ
2
_
1 −
3a
4
r
4
+
2a
2
r
2
_
sin(2θ)
Special points
At the inner hole boundary (r = a), the stress σ
tt
is maximum for θ =
π
2
and the stress
concentration is 3, which means that it is three times the applied stress σ. The stress concen
tration factor (SCF) K
t
is a ratio of the maximum stress and the applied stress and therefore
it is a dimensionless number. It appears not to depend on the hole diameter.
σ
rr
(r = a, θ) = σ
rt
(r = a, θ) = σ
rt
(r, θ = 0) = 0
σ
tt
(r = a, θ =
π
2
) = 3σ
σ
tt
(r = a, θ = 0) = −σ
stress concentration factor K
t
=
σ
max
σ
= 3 []
Stress gradients
When we plot the stress σ
tt
as a function of the distance to the hole, it appears that the stress
gradient is higher when the hole diameter is smaller. For a larger hole there is more volume
of material with a higher than nominal stress level. The chance that there is a defect or ﬂaw
in this volume is higher than in the specimen with the smaller hole. This makes a larger hole
to be more dangerous regarding the occurrence of damage than the smaller hole.
44
Fig. 5.7 : Diﬀerent stress gradient at holes with diﬀerent radius.
5.10 Elliptical hole
Stress components in the vicinity of an elliptical hole can be calculated also. The solution
reveals that at the location of the smallest radius ρ, the stress concentration becomes inﬁnite
when the radius approaches zero, as would be the situation for a real crack.
radius ρ
σ
yy
y
x
a
b
σ
σ
Fig. 5.8 : Elliptical hole loaded with a uniform stress σ in ydirection.
σ
yy
(x = a, y = 0) = σ
_
1 + 2
a
b
_
= σ
_
1 + 2
_
a/ρ
_
≈ 2σ
_
a/ρ
stress concentration factor K
t
= 2
_
a/ρ []
Chapter 6
Crack tip stresses
When a linear crack is loaded in ModeI, II or III, stress components at the crack tip can be
calculated with the Airy stress function method. However, a real crack has zero radius (ρ = 0)
and such a crack tip represents a singularity. Due to this singularity, complex variables and
functions will enter the mathematics.
6.1 Complex plane
The crack tip is considered to be the origin of a complex coordinate system with real axis x
1
and imaginary axis x
2
. The position in this complex plane can be identiﬁed by a complex
number and all variables – Airy function, stresses, strains, displacements – are considered to
be complex functions of this position.
x
1
x
2
r
θ
Fig. 6.1 : Line crack with local coordinate systems originating in the crack tip.
6.1.1 Complex variables
A complex number z has a real (x
1
) and an imaginary (x
2
) part, which are located on the
real and imaginary axis of the complex coordinate system. The conjugate complex number ¯ z
is mirrored with respect to the real axis. The real and imaginary part of z can be expressed
in z and ¯ z.
Introducing unity vectors e
r
and e
i
along the real and imaginary axis, allows the complex
numbers to be represented as a vector z. It is easily seen that multiplying z with the imaginary
constant i results in a counterclockwise rotation over
π
2
of vector z. It is obvious that we can
write : e
i
= ie
r
.
45
46
e
i
e
r
z
¯ z
x
1
θ
r
x
2
Fig. 6.2 : Complex plane and two complex numbers.
z = x
1
+ix
2
= re
iθ
; ¯ z = x
1
−ix
2
= re
−iθ
x
1
=
1
2
(z + ¯ z) ; x
2
=
1
2i
(z − ¯ z) = −
1
2
i(z − ¯ z)
z = x
1
e
r
+x
2
e
i
= x
1
e
r
+x
2
ie
r
= (x
1
+ix
2
)e
r
6.1.2 Complex functions
A scalar function of a complex number z results in a complex number f. We only consider
functions which can be written as the sum of a real (φ) and an imaginary (ζ) part. When
the conjugate complex number ¯ z is substituted in the scalar function, the result is a complex
number
¯
f, which is conjugate to f. It can be shown (see Appendix A) that both the real and
the imaginary part of f (and
¯
f) satisfy the Laplace equation.
f(z) = φ +iζ = φ(x
1
, x
2
) +iζ(x
1
, x
2
) = f
f(¯ z) = φ(x
1
, x
2
) −iζ(x
1
, x
2
) =
¯
f
_
¸
¸
_
¸
¸
_
→
φ =
1
2
{f +
¯
f } ; ζ = −
1
2
i{f −
¯
f }
6.1.3 Laplace operator
The biharmonic equation includes the Laplace operator ∇
2
. In Cartesian coordinates this
operator comprises secondorder derivatives with respect to the Cartesian coordinates x
1
and
x
2
. These derivatives can be transformed to derivatives with respect to the variables z and
¯ z, resulting in a Laplace operator in terms of these variables. In Appendix B a detailed
derivation is presented.
47
complex function g(x
1
, x
2
) = g(z, ¯ z)
Laplacian ∇
2
g =
∂
2
g
∂x
2
1
+
∂
2
g
∂x
2
2
derivatives
∂g
∂x
1
=
∂g
∂z
∂z
∂x
1
+
∂g
∂¯ z
∂¯ z
∂x
1
=
∂g
∂z
+
∂g
∂¯ z
∂
2
g
∂x
2
1
=
∂
2
g
∂z
2
+ 2
∂
2
g
∂z∂¯ z
+
∂
2
g
∂¯ z
2
∂g
∂x
2
=
∂g
∂z
∂z
∂x
2
+
∂g
∂¯ z
∂¯ z
∂x
2
= i
∂g
∂z
−i
∂g
∂¯ z
∂
2
g
∂x
2
2
= −
∂
2
g
∂z
2
+ 2
∂
2
g
∂z∂¯ z
−
∂
2
g
∂¯ z
2
Laplacian ∇
2
g =
∂
2
g
∂x
2
1
+
∂
2
g
∂x
2
2
= 4
∂
2
g
∂z∂¯ z
→ ∇
2
= 4
∂
2
∂z∂¯ z
6.1.4 Biharmonic equation
The Airy stress function ψ is taken to be a function of z and ¯ z and it must be solved from
the biharmonic equation.
Airy function ψ(z, ¯ z)
biharmonic equation ∇
2
_
∇
2
ψ(z, ¯ z)
_
= 0
6.2 Solution of biharmonic equation
According to subsection 6.1.2 (and Appendix A), the solution of a Laplace equation can be
considered to be the real part of a complex function. Thus, from the biharmonic equation it
follows that the Airy function must satisfy a Poisson equation with righthand side the sum
of a complex function and its conjugate. The general solution for ψ contains two unknown
complex functions Ω and ω and there conjugates
¯
Ω and ¯ ω.
∇
2
_
∇
2
ψ(z, ¯ z)
_
= ∇
2
(φ(z, ¯ z)) = 0 → φ = f +
¯
f →
∇
2
ψ = 4
∂
2
ψ
∂z∂¯ z
= φ = f +
¯
f → ψ =
1
2
_
¯ zΩ +z
¯
Ω +ω + ¯ ω
¸
48
6.2.1 Stresses
From the general solution for ψ the stress components σ
11
, σ
22
and σ
12
can be derived and
expressed in Ω, ω,
¯
Ω and ¯ ω. Derivatives w.r.t. z and ¯ z are indicated with ( )
′
. The detailed
derivation can be found in Appendix B.
σ
ij
= σ
ij
(z, ¯ z) = −ψ
,ij
+δ
ij
ψ
,kk
→
σ
11
= −ψ
,11
+ψ
,γγ
= ψ
,22
= Ω
′
+
¯
Ω
′
−
1
2
_
¯ zΩ
′′
+ω
′′
+z
¯
Ω
′′
+ ¯ ω
′′
_
σ
22
= −ψ
,22
+ψ
,γγ
= ψ
,11
= Ω
′
+
¯
Ω
′
+
1
2
_
¯ zΩ
′′
+ω
′′
+z
¯
Ω
′′
+ ¯ ω
′′
_
σ
12
= −ψ
,12
= −
1
2
i
_
¯ zΩ
′′
+ω
′′
−z
¯
Ω
′′
− ¯ ω
′′
_
6.2.2 Displacement
The displacement vector u can be written in components w.r.t. a Cartesian coordinate system
{x
1
, x
2
} with orthonormal base vectors e
1
and e
2
. Considering the coordinate axes to be real
and imaginary axes in a complex plane, we have e
1
= e
r
and e
2
= e
i
= ie
r
. The displacement
vector can thus be considered to be a vector in the direction e
r
, having a length u, which
is a complex number. The real part is the x
1
component and the imaginary part is the
x
2
component.
e
i
x
1
x
2
θ
r
e
2
u
2
u
1
u
e
1
e
r
Fig. 6.3 : Complex displacement.
u = u
1
e
1
+u
2
e
2
= u
1
e
r
+u
2
e
i
= u
1
e
r
+u
2
ie
r
= (u
1
+iu
2
)e
r
= ue
r
→
u = u
1
+iu
2
= u
1
(x
1
, x
2
) +iu
2
(x
1
, x
2
) = u(z, ¯ z)
¯ u = u
1
−iu
2
= ¯ u(z, ¯ z)
Because the displacements u and ¯ u are a function of z and ¯ z, the derivatives w.r.t. z and ¯ z
can be determined. It appears that
∂u
∂¯ z
and
∂¯ u
∂z
can be expressed in strain components only.
49
∂u
∂¯ z
=
∂u
∂x
1
∂x
1
∂¯ z
+
∂u
∂x
2
∂x
2
∂¯ z
=
1
2
_
∂u
∂x
1
+i
∂u
∂x
2
_
=
1
2
_
∂u
1
∂x
1
+i
∂u
2
∂x
1
+i
∂u
1
∂x
2
−
∂u
2
∂x
2
_
=
1
2
(ε
11
−ε
22
+ 2iε
12
)
∂u
∂z
=
∂u
∂x
1
∂x
1
∂z
+
∂u
∂x
2
∂x
2
∂z
=
1
2
_
∂u
∂x
1
−i
∂u
∂x
2
_
=
1
2
_
∂u
1
∂x
1
+i
∂u
2
∂x
1
−i
∂u
1
∂x
2
+
∂u
2
∂x
2
_
=
1
2
_
ε
11
+ε
22
+i
_
∂u
2
∂x
1
−
∂u
1
∂x
2
__
∂¯ u
∂z
=
∂¯ u
∂x
1
∂x
1
∂z
+
∂¯ u
∂x
2
∂x
2
∂z
=
1
2
_
∂¯ u
∂x
1
−i
∂¯ u
∂x
2
_
=
1
2
_
∂u
1
∂x
1
−i
∂u
2
∂x
1
−i
∂u
1
∂x
2
−
∂u
2
∂x
2
_
=
1
2
(ε
11
−ε
22
−2iε
12
)
∂¯ u
∂¯ z
=
∂¯ u
∂x
1
∂x
1
∂¯ z
+
∂¯ u
∂x
2
∂x
2
∂¯ z
=
1
2
_
∂¯ u
∂x
1
+i
∂¯ u
∂x
2
_
=
1
2
_
∂u
1
∂x
1
−i
∂u
2
∂x
1
+i
∂u
1
∂x
2
+
∂u
2
∂x
2
_
=
1
2
_
ε
11
+ε
22
−i
_
∂u
2
∂x
1
−
∂u
1
∂x
2
__
Using Hooke’s law for plane strain, the derivative of u w.r.t. ¯ z can be expressed in the stress
components and thus in the complex functions
¯
Ω and ¯ ω. Integration leads to an expression
for u, comprising an unknown function M(z).
∂u
∂¯ z
=
1
2
1 +ν
E
_
σ
11
−σ
22
+ 2iσ
12
_
= −
1 +ν
E
_
z
¯
Ω
′′
+ ¯ ω
′′
_
u = −
1 +ν
E
_
z
¯
Ω
′
+ ¯ ω
′
+M
_
Using the general solution for u and the conjugate ¯ u, the derivatives of these can be determined
and added. Adding the directly derived derivatives
∂u
∂z
and
∂¯ u
∂¯ z
leads again to an expression
in strain components and through Hooke’s law in the functions Ω
′
and
¯
Ω
′
. From the general
solution, these derivatives can also be calculated, whereupon the integration function M can
be determined.
u = −
1 +ν
E
_
z
¯
Ω
′
+ ¯ ω
′
+M
_
∂u
∂z
= −
1 +ν
E
_
¯
Ω
′
+M
′
¸
¯ u = −
1 +ν
E
_
¯ zΩ
′
+ω
′
+
¯
M
¸
∂¯ u
∂¯ z
= −
1 +ν
E
_
Ω
′
+
¯
M
′
¸
50
∂u
∂z
+
∂¯ u
∂¯ z
= −
1 +ν
E
_
¯
Ω
′
+Ω
′
+M
′
+
¯
M
′
¸
∂u
∂z
+
∂¯ u
∂¯ z
= ε
11
+ε
22
=
1 +ν
E
[(1 −2ν)(σ
11
+σ
22
)]
=
(1 +ν)(1 −2ν)
E
2
_
Ω
′
+
¯
Ω
′
¸
M
′
+
¯
M
′
= −(3 −4ν)
_
¯
Ω
′
+Ω
′
¸
→
M = −(3 −4ν)Ω = −κΩ
u = −
1 +ν
E
_
z
¯
Ω
′
+ ¯ ω
′
−κΩ
_
6.2.3 Choice of complex functions
Two unknown functions Ω and ω have to be determined from boundary conditions. These
two functions are speciﬁed to be of a certain form, however, still containing ﬁve unknown
constants : α, β, γ, δ and λ. Using these functions and their derivatives, the displacement
u(z, ¯ z) can be derived.
The displacement appears to be proportional to r
λ+1
. From a physical point of view,
the displacement always has to be ﬁnite. This argument requires λ to be larger than −1.
Ω = (α +iβ)z
λ+1
= (α +iβ)r
λ+1
e
iθ(λ+1)
ω
′
= (γ +iδ)z
λ+1
= (γ +iδ)r
λ+1
e
iθ(λ+1)
_
→
¯
Ω = (α −iβ)¯ z
λ+1
= (α −iβ)r
λ+1
e
−iθ(λ+1)
¯
Ω
′
= (α −iβ)(λ + 1)¯ z
λ
= (α −iβ)(λ + 1)r
λ
e
−iθλ
¯ ω
′
= (γ −iδ)¯ z
λ+1
= (γ −iδ)r
λ+1
e
−iθ(λ+1)
_
_
_
→
u =
1
2µ
r
λ+1
_
κ(α +iβ)e
iθ(λ+1)
−
(α −iβ)(λ + 1)e
iθ(1−λ)
−(γ −iδ)e
−iθ(λ+1)
_
with µ =
E
2(1 +ν)
displacement ﬁnite → λ > −1
6.2.4 Displacement components
The displacement u is a complex function. Its real and imaginary part, respectively, u
1
and
u
2
, are functions of the variables r and θ and the constants α, β, γ, δ and λ. Euler’s formula
is used to express u in cosine and sine functions.
51
u =
1
2µ
r
λ+1
_
κ(α +iβ)e
iθ(λ+1)
−
(α −iβ)(λ + 1)e
iθ(1−λ)
−(γ −iδ)e
−iθ(λ+1)
_
e
iθ
= cos(θ) +i sin(θ)
_
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
_
u =
1
2µ
r
λ+1
_ _
καcos(θ(λ + 1)) −κβ sin(θ(λ + 1)) −
α(λ + 1) cos(θ(1 −λ)) −β(λ + 1) sin(θ(1 −λ)) −
γ cos(θ(λ + 1)) +δ sin(θ(λ + 1))
_
+
i
_
καsin(θ(λ + 1)) +κβ cos(θ(λ + 1)) −
α(λ + 1) sin(θ(1 −λ)) +β(λ + 1) cos(θ(1 −λ)) +
γ sin(θ(λ + 1)) +δ cos(θ(λ + 1))
_ _
= u
1
+iu
2
6.3 Mode I
For a crack loaded in Mode I, symmetry of the normal displacement and boundary conditions
for crack face stresses, allow the determination of the integration constants.
6.3.1 Displacement
For Mode I loading of the crack, the displacement must be symmetric w.r.t. the x
1
axis, which
is the crack direction. To meet this requirement, the constants β and δ have to be zero. This
can be concluded when studying the displacement components and using sin(−θ) = −sin(θ)
and cos(−θ) = cos(θ). In the functions Ω and ω
′
three unknown constants remain.
u
1
(θ > 0) = u
1
(θ < 0)
u
2
(θ > 0) = −u
2
(θ < 0)
_
→ β = δ = 0 −→
Ω = αz
λ+1
= αr
λ+1
e
i(λ+1)θ
; ω
′
= γz
λ+1
= γr
λ+1
e
i(λ+1)θ
6.3.2 Stress components
The stress components σ
11
, σ
22
and σ
12
can be expressed in the variables r and θ and the
constants α, γ and λ.
52
σ
11
= (λ + 1)
_
αz
λ
+α¯ z
λ
−
1
2
_
αλ¯ zz
λ−1
+γz
λ
+αλz¯ z
λ−1
+γ¯ z
λ
__
σ
22
= (λ + 1)
_
αz
λ
+α¯ z
λ
+
1
2
_
αλ¯ zz
λ−1
+γz
λ
+αλz¯ z
λ−1
+γ¯ z
λ
__
σ
12
= −
1
2
i(λ + 1)
_
αλ¯ zz
λ−1
+γz
λ
−αλz¯ z
λ−1
−γ¯ z
λ
_
with z = re
iθ
; ¯ z = re
−iθ
→
σ
11
= (λ + 1)r
λ
_
αe
iλθ
+αe
−iλθ
−
1
2
_
αλe
i(λ−2)θ
+γe
iλθ
+αλe
−i(λ−2)θ
+γe
−iλθ
__
σ
22
= (λ + 1)r
λ
_
αe
iλθ
+αe
−iλθ
+
1
2
_
αλe
i(λ−2)θ
+γe
iλθ
+αλe
−i(λ−2)θ
+γe
−iλθ
__
σ
12
= −
1
2
i(λ + 1)r
λ
_
αλe
i(λ−2)θ
+γe
iλθ
−αλe
−i(λ−2)θ
−γe
−iλθ
_
with e
iθ
+e
−iθ
= 2 cos(θ) ; e
iθ
−e
−iθ
= 2i sin(θ) →
σ
11
= 2(λ + 1)r
λ
_
αcos(λθ) +
1
2
{αλcos((λ −2)θ) +γ cos(λθ)}
¸
σ
22
= 2(λ + 1)r
λ
_
αcos(λθ) −
1
2
{αλcos((λ −2)θ) +γ cos(λθ)}
¸
σ
12
= (λ + 1)r
λ
[αλsin((λ −2)θ) +γ sin(λθ)]
The boundary condition stating that the crack surface is stressfree, results in two equations
for the unknown constants α and γ. For a nontrivial solution the equations need to be
dependent. This conﬁnes the values of λ to be a series of discrete numbers. For each value of
λ we ﬁnd a ratio between α and γ.
σ
22
(θ = ±π) = σ
12
(θ = ±π) = 0 −→
_
(λ −2) cos(λπ) cos(λπ)
λsin(λπ) sin(λπ)
_ _
α
γ
_
=
_
0
0
_
→
det
_
(λ −2) cos(λπ) cos(λπ)
λsin(λπ) sin(λπ)
_
= −sin(2λπ) = 0 → 2πλ = nπ →
λ = −
1
2
,
n
2
, with n = 0, 1, 2, ..
The stress ﬁeld is a series of terms, one for each value of λ in the range [−
1
2
, 0,
1
2
, 1, · · ·]. The
ﬁrst terms are presented here. Only one unknown constant (γ) remains in these solution.
λ = −
1
2
→ α = 2γ ; λ = 0 → α =
1
2
γ
λ =
1
2
→ α = −2γ ; λ = 1 → α = γ
σ
11
= 2γr
−
1
2
cos(
1
2
θ)
_
1 −sin(
3
2
θ) sin(
1
2
θ)
¸
+· · ·
σ
22
= 2γr
−
1
2
cos(
1
2
θ)
_
1 + sin(
3
2
θ) sin(
1
2
θ)
¸
+· · ·
σ
12
= 2γr
−
1
2
_
cos(
1
2
θ) cos(
3
2
θ) sin(
1
2
θ)
¸
+· · ·
53
6.3.3 Stress intensity factor
The ﬁrst term in the series is the dominant term close to the crack tip. This term represents
the singularity of the stress ﬁeld, because it will become inﬁnite, when approaching the crack
tip. A new constant is introduced in the stress ﬁeld, the socalled Stress Intensity Factor, SIF
or K
I
. It is deﬁned as the limit value of the 22component of the crack tip stress.
K
I
= lim
r→0
_
√
2πr σ
22

θ=0
_
= 2γ
√
2π [ m
1
2
N m
−2
]
6.3.4 Crack tip solution
The stress components are determined for Mode I loading of a straight crack. The SIF K
I
is the only unknown constant in this solution. It can be calculated when more boundary
conditions are available in a situation where geometry and load are speciﬁed in more detail.
The displacement components are also known and for λ = −
1
2
they are proportional to
√
r. For plane stress and plane strain they diﬀer because of the diﬀerent versions of Hooke’s
law.
σ
11
=
K
I
√
2πr
_
cos(
1
2
θ)
_
1 −sin(
1
2
θ) sin(
3
2
θ)
_¸
σ
22
=
K
I
√
2πr
_
cos(
1
2
θ)
_
1 + sin(
1
2
θ) sin(
3
2
θ)
_¸
σ
12
=
K
I
√
2πr
_
cos(
1
2
θ) sin(
1
2
θ) cos(
3
2
θ)
¸
u
1
=
K
I
2µ
_
r
2π
_
cos(
1
2
θ)
_
κ −1 + 2 sin
2
(
1
2
θ)
_¸
u
2
=
K
I
2µ
_
r
2π
_
sin(
1
2
θ)
_
κ + 1 −2 cos
2
(
1
2
θ)
_¸
6.4 Mode II
For a crack loaded in Mode II, antisymmetry of the tangential displacement component and
boundary conditions for the crack face stresses, allow the determination of the integration
constants.
6.4.1 Displacement
For Mode II loading of the crack, the displacement must be antisymmetric w.r.t. the x
1
axis,
which is the crack direction. To meet this requirement, the constants α and γ have to be
zero. This can be concluded from the displacement components, using sin(−θ) = −sin(θ)
and cos(−θ) = cos(θ). In the functions Ω and ω
′
three unknown constants remain.
54
u
1
(θ > 0) = −u
1
(θ < 0)
u
2
(θ > 0) = u
2
(θ < 0)
_
→ α = γ = 0 −→
Ω = iβz
λ+1
= iβr
λ+1
e
i(λ+1)θ
; ω
′
= iδz
λ+1
= iδr
λ+1
e
i(λ+1)θ
6.4.2 Stress intensity factor
The stress intensity factor K
II
is now deﬁned as the limit value of the 12component of the
crack tip stress.
K
II
= lim
r→0
_
√
2πr σ
12

θ=0
_
[ m
1
2
N m
−2
]
6.4.3 Crack tip solution
The stress components are determined for Mode II loading of a straight crack. The SIF K
II
is the only unknown constant in this solution. It can be calculated when more boundary
conditions are available in a situation where geometry and load are speciﬁed in more detail.
The displacement components are also known and for λ = −
1
2
they are proportional to
√
r. For plane stress and plane strain they diﬀer because of the diﬀerent versions of Hooke’s
law.
σ
11
=
K
II
√
2πr
_
−sin(
1
2
θ)
_
2 + cos(
1
2
θ) cos(
3
2
θ)
_¸
σ
22
=
K
II
√
2πr
_
sin(
1
2
θ) cos(
1
2
θ) cos(
3
2
θ)
¸
σ
12
=
K
II
√
2πr
_
cos(
1
2
θ)
_
1 −sin(
1
2
θ) sin(
3
2
θ)
_¸
u
1
=
K
II
2µ
_
r
2π
_
sin(
1
2
θ)
_
κ + 1 + 2 cos
2
(
1
2
θ)
_¸
u
2
=
K
II
2µ
_
r
2π
_
−cos(
1
2
θ)
_
κ −1 −2 sin
2
(
1
2
θ)
_¸
6.5 Mode III
When a crack is loaded in Mode III, only the outofplane displacement u
3
is relevant
55
6.5.1 Laplace equation
The two relevant strain components ε
31
and ε
32
are substututed in Hooke’s law, resulting
in two stress components σ
31
and σ
32
. These are substituted in the equilibrium equation,
leading to a Laplace equation for u
3
.
ε
31
=
1
2
u
3,1
; ε
32
=
1
2
u
3,2
Hooke’s law
_
_
_
→
σ
31
= 2µε
31
= µu
3,1
; σ
32
= 2µε
32
= µu
3,2
equilibrium
_
_
_
→
σ
31,1
+σ
32,2
= µu
3,11
+µu
3,22
= 0 → ∇
2
u
3
= 0
6.5.2 Displacement
The general solution of the Laplace equation is written as the sum of a complex function f
and its conjugate
¯
f. Choosing a speciﬁc function f with three yet unknown constants A,
B and λ, it can again be argumented that the exponent λ must be larger than −1, because
displacements must remain ﬁnite.
general solution u
3
= f +
¯
f
speciﬁc choice f = (A +iB)z
λ+1
→
¯
f = (A −iB)¯ z
λ+1
6.5.3 Stress components
Stress components σ
31
and σ
32
can now be expressed in the variables r and θ and the constants
A, B and λ. The boundary condition that the crack surface is stressfree, results in two
equations for the unknown constants A and B. For a nontrivial solution the equations need
to be dependent. This conﬁnes the values of λ to be a series of discrete numbers. For each
value of λ we ﬁnd a ratio between A and B.
σ
31
= 2(λ + 1)r
λ
{Acos(λθ) −Bsin(λθ)}
σ
32
= −2(λ + 1)r
λ
{Asin(λθ) +Bcos(λθ)}
σ
32
(θ = ±π) = 0 →
_
sin(λπ) cos(λπ)
sin(λπ) −cos(λπ)
_ _
A
B
_
=
_
0
0
_
→
det
_
sin(λπ) cos(λπ)
sin(λπ) −cos(λπ)
_
= −sin(2πλ) = 0 → 2πλ = nπ →
56
λ = −
1
2
,
n
2
, .. with n = 0, 1, 2, ..
crack tip solution λ = −
1
2
→ A = 0 →
σ
31
= Br
−
1
2
{sin(
1
2
θ)} ; σ
32
= −Br
−
1
2
{cos(
1
2
θ)}
6.5.4 Stress intensity factor
The ﬁrst term in the series is the crack tip solution. It determines the singularity of the stress
ﬁeld. The only unknown constant B is replaced by the Stress Intensity Factor K
III
.
K
III
= lim
r→0
_
√
2πr σ
32

θ=0
_
6.5.5 Crack tip solution
The SIF K
III
is the only unknown constant in the stress tip solution. It can be determined
when more boundary conditions are available.
σ
31
=
K
III
√
2πr
_
−sin(
1
2
θ)
¸
; σ
32
=
K
III
√
2πr
_
cos(
1
2
θ)
¸
u
3
=
2K
III
µ
_
r
2π
_
sin(
1
2
θ)
¸
6.6 Crack tip stress (mode I, II, III)
For the three modes of loading, the solution of the biharmonic equation results in a series
expansion both for the displacement components and the stress components. Each discrete
value of the variable λ results in a term of the solution. The dependency of the distance r to
the crack tip is r
λ+1
for the displacements and r
λ
for the stresses. The ﬁrst term is by far
the largest in the vicinity of the crack tip, which is the reason that for the crack tip solution
only the ﬁrst term is considered. The solution is derived for a general crack in a large plate
and is summarized below.
For other crack geometries and loadings, equivalent solutions can be derived, which only
diﬀer in the value of the stress intensity factor K. This K solely determines the amplitude
of the stresses at the crack tip. The dependency of geometry is generally indicated with the
geometry factor β.
57
Mode I Mode II Mode III
σ
τ
τ
σ
τ
τ
Fig. 6.4 : Edge crack under Mode I, II and III loading.
σ
ij
=
K
I
√
2πr
f
Iij
(θ) ; σ
ij
=
K
II
√
2πr
f
IIij
(θ) ; σ
ij
=
K
III
√
2πr
f
IIIij
(θ)
K
I
= β
I
σ
√
πa ; K
II
= β
II
τ
√
πa ; K
III
= β
III
τ
√
πa
6.6.1 Kzone
The question may arise at what distance to the crack tip, displacement and stresses are still
described accurately by the ﬁrst term of the total solution. There is no clear answer to this
question. The Kzone, where the ﬁrst term, whose value is determined by the stress intensity
factor K, is the only important one, depends on geometry and loading, as is illustrated in the
ﬁgure.
D
II I
Kzone : D
D
II
≪D
I
Fig. 6.5 : Kzone at the crack tip in a threepoint bending test.
6.7 SIF for speciﬁed cases
The SIF can be determined when boundary conditions are speciﬁed in more detail. Solu
tions for many cases can be found in literature. Some examples for distributed loading are
presented.
58
W
2a
σ
τ
2a
W
K
I
= σ
√
πa
_
sec
πa
W
_
1/2
K
II
= τ
√
πa small
a
W
a
σ
W
K
I
= σ
√
a
_
1.12
√
π −0.41
a
W
+
18.7
_
a
W
_
2
−38.48
_
a
W
_
3
+
53.85
_
a
W
_
4
_
≈ 1.12σ
√
πa small
a
W
a
σ
W
a
K
I
= σ
√
a
_
1.12
√
π + 0.76
a
W
−
8.48
_
a
W
_
2
+ 27.36
_
a
W
_
3
_
≈ 1.12σ
√
πa
For point loads and a crack loaded by internal pressure, SIF’s are also known from literature.
a
S
P
P/2 P/2
W
K
I
=
PS
BW
3/2
_
2.9
_
a
W
_
1
2
−
4.6
_
a
W
_
3
2
+ 21.8
_
a
W
_
5
2
−
37.6
_
a
W
_
7
2
+ 37.7
_
a
W
_
9
2
_
59
P
a
W
P
K
I
=
P
BW
1/2
_
29.6
_
a
W
_
1
2
−
185.5
_
a
W
_
3
2
+ 655.7
_
a
W
_
5
2
−
1017
_
a
W
_
7
2
+ 638.9
_
a
W
_
9
2
_
p
2a
W
K
I
= p
√
πa
p per unit thickness
6.8 Kbased crack growth criteria
A crack will grow when the crack tip stress exceeds a certain critical value. The stress intensity
factor determines the ”amplitude” of the crack tip stress for a certain geometry and loading
case. We may thus conclude that a crack will grow when K reaches a critical value. This
implies that a crack growth criterion can be formulated, where the stress intensity factor for
a certain situation is compared to this critical value. The value of the stress intensity factor
has to be calculated. The critical value has to be known from experimental measurements.
It is called the Fracture Toughness and denoted as K
c
.
K
I
= K
Ic
; K
II
= K
IIc
; K
III
= K
IIIc
6.9 Relation G−K
The global crack growth criterion was formulated in terms of the energy release rate G and
derived from an energy balance. The local stress growth criterion was formulated in terms
of the stress intensity factor K and derived from crack tip stresses. It may not be surprising
that there is a relation between G and K, which will be derived in this section.
We consider a crack in an ”inﬁnite plate” and focus attention on normal stresses and
openings in the crack plane. First we consider a crack of half length a and then a crack of
half length a+∆a. To close the latter back to length a again, we know which stress is needed
and which opening has to be eliminated. Multiplying normal load and closing displacement,
60
results in the work needed to accomplish this. This work equals the energy which has been
released during opening, the energy release rate G.
x
y
σ
yy
a ∆a
Fig. 6.6 : Central crack opened from length a to length a +∆a.
crack length a
σ
yy
(θ = 0, r = x −a) =
σ
√
a
_
2(x −a)
; u
y
= 0
crack length a +∆a
σ
yy
(θ = π, r = a +∆a −x) = 0
u
y
=
(1 +ν)(κ + 1)
E
σ
√
a +∆a
√
2
√
a +∆a −x
accumulation of elastic energy
∆U = 2B
_
a+∆a
a
1
2
σ
yy
dx u
y
= B
_
a+∆a
a
σ
yy
u
y
dx = Bf(∆a) ∆a →
energy release rate
G =
1
B
lim
∆a→0
_
∆U
∆a
_
= lim
∆a→0
f(∆a) =
(1 +ν)(κ + 1)
4E
σ
2
aπ =
(1 +ν)(κ + 1)
4E
K
2
I
plane stress G =
K
2
I
E
plane strain G = (1 −ν
2
)
K
2
I
E
A relation between G and K is also available for multimode loading and can be derived by
superposition.
61
G =
1
E
_
c
1
K
2
I
+c
2
K
2
II
+c
3
K
2
III
_
plane stress G =
1
E
(K
2
I
+K
2
II
)
plane strain G =
(1 −ν
2
)
E
(K
2
I
+K
2
II
) +
(1 +ν)
E
K
2
III
6.10 The critical SIF value
In the local crack growth criterion, the actual value K of the stress intensity factor is compared
to a critical value K
c
. This has to be measured according to an elaborate experimental
procedure and is then considered to be a material constant. The actual procedure can be
found in literature [25] and is normalized.
When K
Ic
would be measured using a large plate with a central crack of length 2a, the
critical value would appear to be dependent on the plate thickness. This can be explained
later – we have to consider the plastic crack tip zone for this –. It appears that the critical
thickness B
c
, from which a constant value for K
Ic
results, is very large. This means that such
an experimental setup would be unfeasible and therefore other experiments are designed and
normalized.
σ
2a
B
σ
K
I
B
c
B
K
Ic
Fig. 6.7 : Central crack in a plate, loaded with Mode I stress σ (left).
Dependence of critical Kvalue on the plate thickness B (right).
K
Ic
= σ
c
√
πa ; B
c
= 2.5
_
K
Ic
σ
y
_
2
62
6.10.1 K
Ic
values
The table lists values for the fracture toughness K
Ic
for various materials. They are taken
from [25].
Material σ
v
[MPa] K
Ic
[MPa
√
m ]
steel, 300 maraging 1669 93.4
steel, 350 maraging 2241 38.5
steel, D6AC 1496 66.0
steel, AISI 4340 1827 47.3
steel, A533B reactor 345 197.8
steel, carbon 241 219.8
Al 2014T4 448 28.6
Al 2024T3 393 34.1
Al 7075T651 545 29.7
Al 7079T651 469 33.0
Ti 6Al4V 1103 38.5
Ti 6Al6V2Sn 1083 37.4
Ti 4Al4Mo2Sn0.5Si 945 70.3
Chapter 7
Multimode crack loading
In practical situations, a crack is mostly subjected to a combined Mode I and Mode II loading.
In the ﬁgure the load is visualized as an edge load on a square material volume element with
edges parallel and perpendicular to the crack. A tensile load parallel to the crack can be added
because it has no inﬂuence on the crack tip stress. Each random load can be transformed
to this loading situation by proper transformation of the edge loads. The crack tip stresses
can be determined as the superposition of the stress components due to separate Mode I and
Mode II loadings.
Mode II Mode I
Mode I +II
Mode I +II
Fig. 7.1 : Central crack loaded in Mode I, Mode II and a combination of Mode I and II.
7.1 Stress component transformation
We consider a plane with unity normal vector n and stress vector p. Components of n and
p with respect to the base {e
1
, e
2
} are stored in columns n
˜
and p
˜
. Another base {e
∗
1
, e
∗
2
}
is rotated anticlockwise over an angle θ w.r.t. {e
1
, e
2
}. Components of n and p w.r.t. this
63
64
base are stored in columns n
˜
∗
and p
˜
∗
. Using the relation between {e
∗
1
, e
∗
2
} and {e
1
, e
2
}, a
transformation matrix T can be determined which relates the columns n
˜
∗
to n
˜
, and p
˜
∗
to p
˜
.
n
p
e
∗
2
θ
(b)
e
2
e
1
e
∗
1
Fig. 7.2 : Material plane with unit normal vector and stress vector.
e
∗
1
= cos(θ)e
1
+ sin(θ)e
2
= ce
1
+se
2
e
∗
2
= −sin(θ)e
1
+ cos(θ)e
2
= −se
1
+ce
2
p = p
1
e
1
+p
2
e
2
= p
∗
1
e
∗
1
+p
∗
2
e
∗
2
→
_
p
1
p
2
_
=
_
c −s
s c
_ _
p
∗
1
p
∗
2
_
→
_
p
∗
1
p
∗
2
_
=
_
c s
−s c
_ _
p
1
p
2
_
→
p
˜
= T p
˜
∗
→ p
˜
∗
= T
T
p
˜
idem : n
˜
∗
= T
T
n
˜
With respect to the basis {e
1
, e
2
}, the stress vector components p
˜
can be calculated from the
Cauchy stress matrix σ and the components n
˜
of the normal vector. Using the transformation
matrix T, the matrix σ
∗
w.r.t. the basis {e
∗
1
, e
∗
2
} can be calculated from σ.
p
˜
= σn
˜
→
T p
˜
∗
= σ Tn
˜
∗
→ p
˜
∗
= T
T
σ T n
˜
∗
= σ
∗
n
˜
∗
→
σ
∗
= T
T
σ T → σ = T σ
∗
T
T
_
σ
∗
11
σ
∗
12
σ
∗
21
σ
∗
22
_
=
_
c s
−s c
_ _
σ
11
σ
12
σ
21
σ
22
_ _
c −s
s c
_
=
_
c s
−s c
_ _
cσ
11
+sσ
12
−sσ
11
+cσ
12
cσ
21
+sσ
22
−sσ
21
+cσ
22
_
=
_
¸
¸
_
c
2
σ
11
+ 2csσ
12
+s
2
σ
22
−csσ
11
+ (c
2
−s
2
)σ
12
+csσ
22
−csσ
11
+ (c
2
−s
2
)σ
12
+csσ
22
s
2
σ
11
−2csσ
12
+c
2
σ
22
_
¸
¸
_
65
Stress component transformation can also be used to determine stress components in a cylin
drical coordinate system. These will be needed later in crack growth direction criteria. With
the ﬁgure below, it is easily understood that cylindrical stress components can be derived
from Cartesian components by transformation of the stress matrix.
σ
rr
σ
rt
σ
tt
σ
xy
r
θ
e
1
e
2
e
r
e
t
σ
xx
σ
yy
Fig. 7.3 : Cartesian and cylindrical stress ’cubes’ at the crack tip.
e
r
= ce
1
+se
2
; e
t
= −se
1
+ce
2
_
σ
rr
σ
rt
σ
tr
σ
tt
_
=
_
c s
−s c
_ _
σ
xx
σ
xy
σ
xy
σ
yy
_ _
c −s
s c
_
=
_
¸
¸
_
c
2
σ
xx
+ 2csσ
xy
+s
2
σ
yy
−csσ
xx
+ (c
2
−s
2
)σ
xy
+csσ
yy
−csσ
xx
+ (c
2
−s
2
)σ
xy
+csσ
yy
s
2
σ
xx
−2csσ
xy
+c
2
σ
yy
_
¸
¸
_
Crack tip stresses : Cartesian
The Cartesian crack tip stresses have been derived before and are summarized below.
σ
xx
σ
yy
σ
xy
Fig. 7.4 : Cartesian stress cube at the crack tip.
66
σ
xx
=
K
I
√
2πr
f
Ixx
(θ) +
K
II
√
2πr
f
IIxx
(θ)
σ
yy
=
K
I
√
2πr
f
Iyy
(θ) +
K
II
√
2πr
f
IIyy
(θ)
σ
xy
=
K
I
√
2πr
f
Ixy
(θ) +
K
II
√
2πr
f
IIxy
(θ)
f
Ixx
(θ) = cos(
θ
2
)
_
1 −sin(
θ
2
) sin(
3θ
2
)
¸
f
IIxx
(θ) = −sin(
θ
2
)
_
2 + cos(
θ
2
) cos(
3θ
2
)
¸
f
Iyy
(θ) = cos(
θ
2
)
_
1 + sin(
θ
2
) sin(
3θ
2
)
¸
f
IIyy
(θ) = sin(
θ
2
) cos(
θ
2
) cos(
3θ
2
)
f
Ixy
(θ) = sin(
θ
2
) cos(
θ
2
) cos(
3θ
2
)
f
IIxy
(θ) = cos(
θ
2
)
_
1 −sin(
θ
2
) sin(
3θ
2
)
¸
Substitution in the transformation relation, results in cylindrical crack tip stresses. The
functions f
rr
, f
tt
and f
rt
can be derived, using some trigonometry.
σ
tt θ
σ
rr
σ
rt
Fig. 7.5 : Cylindrical stress ’cube’ at the crack tip.
σ
rr
=
K
I
√
2πr
f
Irr
(θ) +
K
II
√
2πr
f
IIrr
(θ)
σ
tt
=
K
I
√
2πr
f
Itt
(θ) +
K
II
√
2πr
f
IItt
(θ)
σ
rt
=
K
I
√
2πr
f
Irt
(θ) +
K
II
√
2πr
f
IIrt
(θ)
f
Irr
(θ) =
_
5
4
cos(
θ
2
) −
1
4
cos(
3θ
2
)
¸
f
IIrr
(θ) =
_
−
5
4
sin(
θ
2
) +
3
4
sin(
3θ
2
)
¸
f
Itt
(θ) =
_
3
4
cos(
θ
2
) +
1
4
cos(
3θ
2
)
¸
f
IItt
(θ) =
_
−
3
4
sin(
θ
2
) −
3
4
sin(
3θ
2
)
¸
f
Irt
(θ) =
_
1
4
sin(
θ
2
) +
1
4
sin(
3θ
2
)
¸
f
IIrt
(θ) =
_
1
4
cos(
θ
2
) +
3
4
cos(
3θ
2
)
¸
67
7.2 Multimode load
For a crack with angle θ w.r.t. the edge of the material volume element with edge loads
represented by σ, the Mode I and Mode II loading is represented by the edge loads σ
∗
on a rotated material volume element. The crack tip stresses of the combined loading are
determined by superposition. The load parallel to the crack (σ
∗
11
) is assumed to have no
inﬂuence on the crack tip stress state. Note that in the next cases the load is indicated
by σ
ij
and the crack tip stresses by s
ij
.
2a σ
11
σ
22
σ
12
σ
∗
11
σ
∗
22
σ
∗
12
θ
e
1
e
2
σ
∗
12
σ
∗
12
σ
∗
22
e
∗
2
e
∗
1
2a
Fig. 7.6 : Transformation of multimode load into Mode I and II loads.
load
_
σ
∗
11
σ
∗
12
σ
∗
21
σ
∗
22
_
=
_
¸
¸
_
c
2
σ
11
+ 2csσ
12
+s
2
σ
22
−csσ
11
+ (c
2
−s
2
)σ
12
+csσ
22
−csσ
11
+ (c
2
−s
2
)σ
12
+csσ
22
s
2
σ
11
−2csσ
12
+c
2
σ
22
_
¸
¸
_
crack tip stresses s
ij
=
K
I
√
2πr
f
Iij
(θ) +
K
II
√
2πr
f
IIij
(θ)
with K
I
= β σ
∗
22
√
πa ; K
II
= γ σ
∗
12
√
πa
Example multimode load
As an example we consider the biaxial load on a crack which orientation has an angle θ with
the horizontal direction.
θ
σ
kσ
σ
∗
22
σ
∗
11
σ
∗
12
σ
∗
12
2a 2a
Fig. 7.7 : Transformation of multimode load into Mode I and II loads.
68
load
σ
∗
11
= c
2
σ
11
+ 2csσ
12
+s
2
σ
22
= c
2
kσ +s
2
σ
σ
∗
22
= s
2
σ
11
−2csσ
12
+c
2
σ
22
= s
2
kσ +c
2
σ
σ
∗
12
= −csσ
11
+ (c
2
−s
2
)σ
12
+csσ
22
= cs(1 −k)σ
crack tip stresses
s
ij
=
K
I
√
2πr
f
Iij
(θ) +
K
II
√
2πr
f
IIij
(θ)
stress intensity factors
K
I
= β
I
σ
∗
22
√
πa = β
I
(s
2
k +c
2
)σ
√
πa
K
II
= β
II
σ
∗
12
√
πa = β
II
cs(1 −k)σ
√
πa
Example multimode load
The results from the example above, can be used directly in the situation of a crack through
the whole thickness of the wall of a thinwalled tube. The crack has an angle θ w.r.t. the
axial direction.
θ
p
2a
R
t
σ
∗
22
σ
∗
12
σ
∗
11
σ
a
σ
t
Fig. 7.8 : Crack in tube wall at an angle θ with the tube axis.
σ
t
=
pR
t
= σ ; σ
a
=
pR
2t
=
1
2
σ → k =
1
2
σ
∗
22
= s
2 1
2
σ +c
2
σ ; σ
∗
12
= cs(1 −
1
2
)σ =
1
2
cs σ
K
I
= σ
∗
22
√
πa = (
1
2
s
2
+c
2
)σ
√
πa = (
1
2
s
2
+c
2
)
pR
t
√
πa
K
II
= σ
∗
12
√
πa =
1
2
cs σ =
1
2
cs
pR
t
√
πa
69
7.3 Crack growth direction
Crack growth criteria are used to check whether a crack will grow at a given external load.
When the crack is loaded only in mode I, it will be expected that the crack tip will move in
the crack plane. However, in multimode loading, the crack growth direction is not so easy
to predict.
Two crack growth direction criteria will be described in the following :
• maximum tangential stress (MTS) criterion,
• strain energy density (SED) criterion.
For special loading cases it will be possible to calculate the crack growth direction by analytical
means. More general cases are explored with numerical techniques.
For both criteria, we have to use stress and displacement components in a cylindrical
coordinate system.
7.3.1 Maximum tangential stress criterion
The ﬁrst crack growth direction criterion to be discussed here is the maximum tangential
stress criterion, which was suggested and published by Erdogan and Sih in 1963 [18]. The
hypothesis is that a crack tip will move in the direction of the point where the highest value
of the tangential stress component is found. This can be expressed mathematically by ﬁrst
and second derivatives of the tangential stress component w.r.t. the angle θ.
σ
tt θ
σ
rr
σ
rt
Fig. 7.9 : Cylindrical stress components at the crack tip.
Hypothesis : crack growth towards local maximum of σ
tt
∂σ
tt
∂θ
= 0 and
∂
2
σ
tt
∂θ
2
< 0 → θ
c
σ
tt
(θ = θ
c
) = σ
tt
(θ = 0) =
K
Ic
√
2πr
→ crack growth
70
Because the function σ
tt
(θ) is known, the relations can be elaborated.
∂σ
tt
∂θ
= 0 →
3
2
K
I
√
2πr
_
−
1
4
sin(
θ
2
) −
1
4
sin(
3θ
2
)
¸
+
3
2
K
II
√
2πr
_
−
1
4
cos(
θ
2
) −
3
4
cos(
3θ
2
)
¸
= 0 →
K
I
sin(θ) +K
II
{3 cos(θ) −1} = 0
∂
2
σ
tt
∂θ
2
< 0 →
3
4
K
I
√
2πr
_
−
1
4
cos(
θ
2
) −
3
4
cos(
3θ
2
)
¸
+
3
4
K
II
√
2πr
_
1
4
sin(
θ
2
) +
9
4
sin(
3θ
2
)
¸
< 0
σ
tt
(θ = θ
c
) =
K
Ic
√
2πr
→
1
4
K
I
K
Ic
_
3 cos(
θc
2
) + cos(
3θc
2
)
¸
+
1
4
K
II
K
Ic
_
−3 sin(
θc
2
) −3 sin(
3θc
2
)
¸
= 1
Mode I load
For mode I loading of the crack, it is expected that the crack tip will advance in the plane of
the crack. Indeed this obviously results from solving the relations for θ.
∂σ
tt
∂θ
= K
I
sin(θ) = 0 → θ
c
= 0
∂
2
σ
tt
∂θ
2
¸
¸
¸
¸
θc
< 0
σ
tt
(θ
c
) =
K
Ic
√
2πr
→ K
I
= K
Ic
Mode II load
For pure mode II loading of the crack, the crack growth direction has to be determined
from the relations for θ. The requirement that the ﬁrst derivative of σ
tt
(θ) is zero, leads
to an equation, which results in two candidate values for θ
c
. Only for the negative value
θ
c
= −70.6
o
, the second requirement is satisﬁed, so that we can conclude that the crack tip
will move in that direction. The critical value of the mode II stress intensity factor can be
expressed in the critical value of the mode I SIF.
∂σ
tt
∂θ
= K
II
(3 cos(θ
c
) −1) = 0 → θ
c
= ±arccos(
1
3
) = ±70.6
o
∂
2
σ
tt
∂θ
2
¸
¸
¸
¸
θc
< 0 → θ
c
= −70.6
o
σ
tt
(θ
c
) =
K
Ic
√
2πr
→ K
IIc
=
_
3
4
K
Ic
71
τ
θ
c
τ
Fig. 7.10 : Crack growth direction for Mode II loading.
Multimode load
Determining the crack growth direction for multimode loading is only possible using nu
merical calculations. The relations which have to be satisﬁed for crack growth in a certain
direction are written is terms of the parameter
K
I
K
Ic
.
K
I
[−sin(
θ
2
) −sin(
3θ
2
)] +K
II
[−cos(
θ
2
) −3 cos(
3θ
2
)] = 0
K
I
[−cos(
θ
2
) −3 cos(
3θ
2
)] +K
II
[sin(
θ
2
) + 9 sin(
3θ
2
)] < 0
K
I
[3 cos(
θ
2
) + cos(
3θ
2
)] +K
II
[−3 sin(
θ
2
) −3 sin(
3θ
2
)] = 4K
Ic
−K
I
f
1
−K
II
f
2
= 0
−K
I
f
2
+K
II
f
3
< 0
K
I
f
4
−3K
II
f
1
= 4K
Ic
_
¸
¸
¸
¸
_
¸
¸
¸
¸
_
→
−
_
K
I
K
Ic
_
f
1
−
_
K
II
K
Ic
_
f
2
= 0
−
_
K
I
K
Ic
_
f
2
+
_
K
II
K
Ic
_
f
3
< 0
_
K
I
K
Ic
_
f
4
−3
_
K
II
K
Ic
_
f
1
= 4
_
¸
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
¸
_
For a range of values 0 ≤
K
I
K
Ic
≤ 1 a range of values for the crack growth direction angle
0 ≤ θ ≤ 90
o
is evaluated to decide if the relations are satisﬁed. In the ﬁrst one
K
II
K
Ic
is plotted
as a function of
K
I
K
Ic
and it can be used to determine which combination leads to crack growth.
From the second plot the crack growth angle θ
c
can then be determined.
0 0.2 0.4 0.6 0.8 1
−70
−60
−50
−40
−30
−20
−10
0
K
I
/K
Ic
θ
c
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
K
I
/K
Ic
K
I
I
/
K
I
c
Fig. 7.11 : Crack growth direction and Mode ratio for multimode loading.
72
7.3.2 Strain energy density (SED) criterion
The strain energy density U
i
is the stored elastic energy per unit of volume. For linear elastic
material behavior this speciﬁc energy is easily calculated in the crack tip region and appears
to be inverse proportional to the distance to the crack tip. The strain energy density factor
S is now deﬁned as the product rU
i
, as such being independent of r.
The strain energy density criterion is proposed and published by Sih in 1973 [61] and
states that at crack growth, the crack tip will move toward the point where S is minimum.
U
i
= Strain Energy Density (Function) =
_
ε
ij
0
σ
ij
dε
ij
S = Strain Energy Density Factor = rU
i
= S(K
I
, K
II
, θ)
Hypothesis : crack growth towards local minimum of SED
∂S
∂θ
= 0 and
∂
2
S
∂θ
2
> 0 → θ
c
S(θ = θ
c
) = S(θ = 0, pl.strain) = S
c
→ crack growth
For the case of linear elastic material behavior, the strain energy density function can be
expressed in the stress components and the material parameters E and ν. At the crack tip
these stress components are known functions of r, θ, K
I
and K
II
. The strain energy density
factor is than also known and dependent on θ. K
I
and K
II
.
The requirements for the crack growth direction result in two relations, from which θ
c
can be determined.
U
i
=
1
2E
(σ
2
xx
+σ
2
yy
+σ
2
zz
) −
ν
E
(σ
xx
σ
yy
+σ
yy
σ
zz
+σ
zz
σ
xx
) +
1
2G
(σ
2
xy
+σ
2
yz
+σ
2
zx
)
σ
xx
=
K
I
√
2πr
cos(
θ
2
)
_
1 −sin(
θ
2
) sin(
3θ
2
)
¸
−
K
II
√
2πr
sin(
θ
2
)
_
2 + cos(
θ
2
) cos(
3θ
2
)
¸
σ
yy
=
K
I
√
2πr
cos(
θ
2
)
_
1 + sin(
θ
2
) sin(
3θ
2
)
¸
+
K
II
√
2πr
sin(
θ
2
) cos(
θ
2
) cos(
3θ
2
)
σ
xy
=
K
I
√
2πr
sin(
θ
2
) cos(
θ
2
) cos(
3θ
2
) +
K
II
√
2πr
cos(
θ
2
)
_
1 −sin(
θ
2
) sin(
3θ
2
)
¸
73
S = rU
i
= S(K
I
, K
II
, θ) = a
11
k
2
I
+ 2a
12
k
I
k
II
+a
22
k
2
II
with a
11
=
1
16G
(1 + cos(θ))(κ −cos(θ))
a
12
=
1
16G
sin(θ){2 cos(θ) −(κ −1)}
a
22
=
1
16G
{(κ + 1)(1 −cos(θ)) + (1 + cos(θ))(3 cos(θ) −1)}
k
i
= K
i
/
√
π
Mode I load
For Mode I loading we have K
II
= 0 and K
I
= σ
√
πa. The ﬁrst requirement results in two
possible values of θ
c
. The second requirement is only satisﬁed for θ
c
= 0, because 1 ≤ κ ≤ 3.
The obvious crack growth direction is θ
c
= 0. Crack growth will take place if the value of
S(θ
c
) equals the critical value S
c
, which is the value for plane strain.
S = a
11
k
2
I
=
σ
2
a
16G
{1 + cos(θ)}{κ −cos(θ)}
∂S
∂θ
= sin(θ){2 cos(θ) −(κ −1)} = 0 →
θ
c
= 0 or arccos
_
1
2
(κ −1)
_
∂
2
S
∂θ
2
= 2 cos(2θ) −(κ −1) cos(θ) > 0 → θ
c
= 0
S(θ
c
) =
σ
2
a
16G
{2}{κ −1} =
σ
2
a
8G
(κ −1)
S
c
= S(θ
c
, pl.strain) =
(1 +ν)(1 −2ν)
2πE
K
2
Ic
Mode II load
For Mode II loading we have K
I
= 0 and K
II
= τ
√
πa. The two requirements lead to two
possible crack growth directions. The result diﬀers for plane stress and plane strain. For
ν = 0, leading to κ = 3, the direction from SED is the same as the one from MTS. Crack
growth will occur if the minimum Svalue equals the critical value S
c
, which was determined
in pure Mode I loading. From this value the critical shear stress can be derived.
S = a
22
k
2
II
=
τ
2
a
16G
[(κ + 1){1 −cos(θ)} +{1 + cos(θ)}{3 cos(θ) −1}]
∂S
∂θ
= sin(θ) [−6 cos(θ) + (κ −1)] = 0
∂
2
S
∂θ
2
= 6 −cos
2
(θ) + (κ −1) cos(θ) > 0
_
¸
¸
_
¸
¸
_
→
θ
c
= ±arccos
_
1
6
(κ −1)
_
S(θ
c
) =
τ
2
a
16G
{
1
12
(−κ
2
+ 14κ −1)}
S(θ
c
) = S
c
→ τ
c
=
1
√
a
_
192GS
c
−κ
2
+ 14κ −1
74
Multimode load
It is again not possible to determine the crack growth direction for a general multimode
loading. Numerically it can be done, however.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
−90
−80
−70
−60
−50
−40
−30
−20
−10
0
10
K
I
/K
Ic
θ
c
ν=0.0.75
ν=0.1
ν=0.2
ν=0.3
ν=0.4
ν=0.495
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
K
I
/K
Ic
K
II /
K
Ic
ν=0.0.75
ν=0.1
ν=0.2
ν=0.3
ν=0.4
ν=0.495
Fig. 7.12 : Crack growth direction and Mode ratio for multimode loading.
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
−90
−80
−70
−60
−50
−40
−30
−20
−10
0
10
K
I
/K
Ic
θ
c
ν=0.0.75
ν=0.1
ν=0.2
ν=0.3
ν=0.4
ν=0.495
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
K
I
/K
Ic
K
II /
K
Ic
ν=0.0.75
ν=0.1
ν=0.2
ν=0.3
ν=0.4
ν=0.495
Fig. 7.13 : Crack growth direction and Mode ratio for multimode loading.
The ﬁgure below shows a crack oriented at an angle β w.r.t. the tensile load σ. For this
situation the scaled stress intensity factors k
I
and k
II
can be calculated, from which follows
the expression for S as a function of β and θ. The requirements according to the SED criterion
can be used to test a sequence of θ
c
values for a sequence of β values and a plot can be made.
The ﬁgure shows the result for plane strain.
k
I
= σ
√
a sin
2
(β) ; k
II
= σ
√
a sin(β) cos(β)
S = σ
2
a sin
2
(β)
_
a
11
sin
2
(β) + 2a
12
sin(β) cos(β) +a
22
cos
2
(β)
_
∂S
∂θ
= (κ −1) sin(θ
c
−2β) −2 sin{2(θ
c
−β)} −sin(2θ
c
) = 0
∂
2
S
∂θ
2
= (κ −1) cos(θ
c
−2β) −4 cos{2(θ
c
−β)} −2 cos(2θ
c
) > 0
75
2a
β
θ
c
σ
σ
β
ν = 0.1
ν = 0
ν = 0.5
90
−θ
c
80
70
60
50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90
Fig. 7.14 : Crack growth directions for multimode loading.
76
Chapter 8
Dynamic fracture mechanics
When a crack is subjected to impact loading, the crack tip stresses and strains can only
be calculated by taking into account dynamic eﬀects like material acceleration and solving
the equations of motion. This is the subject of Dynamic Fracture Mechanics (DFM) and
understandably involves rather sophisticated mathematics.
Another subject, studied in DFM, is the behavior of a crack, which grows at high rate
under (quasi)static loading after reaching its critical length. In this chapter an approximate
value for the crack tip speed is derived, using the global energy balance, taking into account
the kinetic energy term. This approach is referred to as the kinetic approach. In DFM a
socalled static approach is also used to study the behavior of cracks, subjected to thermal
shock loading.
8.1 Crack growth rate
To calculate the crack tip speed or crack growth rate, the general energy balance is used
as a starting point. It is elaborated for a crack with length 2a in an inﬁnitely large plate
with uniform and constant thickness B. The crack is subjected to a Mode I load σ, resulting
from an external load applied at edges far away from the crack. This implies that dU
e
= 0
when the crack length changes. It is assumed that the presence of the crack has led to a
complete release of elastic energy U
i
in the elliptical area, indicated in the ﬁgure. Dissipation
is assumed to be zero, so all available energy is going to be transformed into surface energy
U
a
and kinetic energy U
k
.
dU
e
da
−
dU
i
da
=
dU
a
da
+
dU
d
da
+
dU
k
da
77
78
2a
a
x
σ
σ
y
thickness B
Fig. 8.1 : Central crack in large plate with unloaded elliptical area.
dU
e
da
= 0 ;
dU
d
da
= 0
U
a
= 4aBγ →
dU
a
da
= 4γB
U
i
= 2πa
2
B
1
2
σ
2
E
→ −
dU
i
da
=
2πaBσ
2
E
Kinetic energy
The total kinetic energy of the plate is the integral over its volume of the speciﬁc kinetic
energy
1
2
ρ( ˙ u
2
x
+ ˙ u
2
y
), where ρ is the material density and ˙ u
x
and ˙ u
y
the material velocity
components.
When the crack opens, the material velocity perpendicular to the crack faces ( ˙ u
y
) is
much higher than the velocity in the crack plane ( ˙ u
x
). To allow further derivations, it is
assumed that the crack tip speed s is independent of the (half) crack length a.
The displacement u
y
of the crack face (θ = π) is available from the analytical solution
for a static crack, here assumed to be in a state of plane stress. The kinetic energy change at
crack growth can now be calculated by integration. However, the integral cannot be calculated
analytically. Its value is indicated as ak(a) and will be considered later.
U
k
=
1
2
ρB
_
Ω
( ˙ u
2
x
+ ˙ u
2
y
) dxdy
material velocity ˙ u
x
≪ ˙ u
y
=
du
y
dt
=
du
y
da
da
dt
=
du
y
da
s
_
¸
¸
_
¸
¸
_
U
k
=
1
2
ρs
2
B
_
Ω
_
du
y
da
_
2
dxdy
assumption
ds
da
= 0
_
¸
¸
_
¸
¸
_
79
dU
k
da
=
1
2
ρs
2
B
_
Ω
d
da
_
du
y
da
_
2
dxdy
u
y
= 2
√
2
σ
E
_
a
2
−ax →
du
y
da
=
√
2
σ
E
2a −x
√
a
2
−ax
_
¸
¸
¸
_
¸
¸
¸
_
dU
k
da
= ρs
2
B
_
σ
E
_
2
a
_
Ω
1
a
3
x
2
(x −2a)
(a −x)
2
dxdy = ρs
2
B
_
σ
E
_
2
a k(a)
Substitution of all results in the energy balance leads to an expression for the crack tip speed
s where the critical crack length a
c
and the sound speed c can be introduced.
The earlier assumption that s is independent of a does not appear to hold, so the result
must be an approximation. More approximation is involved in the calculation of k(a), which
is done numerically. Roberts and Wells did this in 1954 [56] and came up with k = 43.51.
As can be seen from the expression for s, the crack will grow at high speed, so that the
actual (half) crack length a will be much larger than the critical crack length a
c
. This leads
to the ﬁnal approximation for the crack tip speed.
2πaσ
2
E
= 4γ +ρs
2
_
σ
E
_
2
ak →
s =
_
E
ρ
_
1
2
_
2π
k
_
1
2
_
1 −
2γE
πaσ
2
_
1
2
_
2π
k
≈ 0.38 ; a
c
=
2γE
πσ
2
; c =
¸
E
ρ
_
¸
¸
¸
¸
¸
¸
_
¸
¸
¸
¸
¸
¸
_
s = 0.38 c
_
1 −
a
c
a
_
1
2
a ≫a
c
_
¸
¸
_
¸
¸
_
→ s ≈ 0.38 c
For some materials the table lists values for Young’s modulus E, density ρ and calculated
sound speed c. The measured crack tip speed s is also listed. The ratio s/c takes values
between 0.2 and 0.4, so the approximation 0.38 is well within this range. Mostly the experi
mental ratio is found to be smaller, due to the dissipation during crack growth.
steel copper aluminum glass rubber
E [GPa] 210 120 70 70 20
ρ [kg/m
2
] 7800 8900 2700 2500 900
ν 0.29 0.34 0.34 0.25 0.5
c [m/sec] 5190 3670 5090 5300 46
s [m/sec] 1500 2000
s/c 0.29 0.38
80
0.2 <
s
c
< 0.4
8.2 Elastic wave speeds
In the derivation of the crack tip speed s, the sound speed c was introduced, which is the
speed of a wave front with elastic elongational strain, also indicated as C
0
. In DFM more
sophisticated analyzes of crack growth rate use other deformation modes, e.g. dilatation (vol
ume change) and shear, with wave front speeds C
1
and C
2
, respectively. Frequently used is
the Rayleigh velocity C
R
of wave fronts traveling at the surface of a material.
C
0
= elongational wave speed =
¸
E
ρ
C
1
= dilatational wave speed =
_
κ + 1
κ −1
_
µ
ρ
C
2
= shear wave speed =
_
µ
ρ
C
R
= Rayleigh velocity = 0.54 C
0
´a 0.62 C
0
Corrections
More elaborate derivations of the crack tip speed have been published. Dulancy & Brace
(1960) derived an expression without assuming s to be independent of a. Freund (1972)
related s to the Rayleigh velocity C
R
.
Dulancy & Brace (1960) s = 0.38 C
0
_
1 −
a
c
a
_
Freund (1972) s = C
R
_
1 −
a
c
a
_
8.3 Crack tip stress
When a crack tip moves at high speed through a material, the elastic stresses σ
Dij
at the
crack tip can be calculated [70]. The ﬁrst term in this stress ﬁeld appears to be proportional
to K
D
/
√
2πr, with r the distance to the crack tip and K
D
the Dynamic Stress Intensity
Factor. The stress components are, just as in the static case, a function of the angle θ w.r.t.
the crack plane. In this dynamic case, however, they also depend on material parameters and
the crack tip speed.
σ
Dij
=
K
D
√
2πr
f
ij
(θ, r, s, E, ν)
81
8.3.1 Crack branching
Cylindrical stress components for Mode I loading can be derived by transformation as has
been done for the static loading case. The ratio of the tangential stress σ
IDtt
(θ) and the
tangential stress in the crack plane (θ = 0), can be calculated as a function of θ. The
graphical representation shows diﬀerent curves for various values of the crack tip speed s.
For s/C
R
≈ 0.87 the curve has a maximum value at θ ≈ π/4. According to the Maximum
Tangential Stress criterion, the crack will move out of the crack plane. This crack branching
will occur repeatedly, when the crack runs through the material, and will soon lead to complete
material failure.
σ
Dij
=
K
ID
√
2πr
f
ij
(θ, r, s, E, ν)
volgens MTS
max
θ
π
π
2
crack branching
s
c
R
σ
tt
0
0.87
1 0.6
σ
Dtt
(θ)
σ
Dtt
(θ = 0)
0.9
Fig. 8.2 : Tangential stress ratio as a function of crack growth direction, illustrating crack
branching.
8.3.2 Fast fracture and crack arrest
A crack, which runs through a material, can reach a location, where it is subjected to diﬀerent
loading conditions, which may result in crack arrest. This will occur when K
D
is below a
critical value K
A
, being the minimal value of a measured Dynamic Fracture Toughness K
Dc
,
which is a function of crack tip speed s and temperature T.
K
D
≥ K
Dc
(s, T) → crack growth
K
D
< min
0<s<C
R
K
Dc
(s, T) = K
A
→ crack arrest
82
8.4 Experiments
In DFM experiments are done to measure crack tip speeds, using instrumented DCB speci
mens, sometimes in combination with high speed photography (typically 10
6
frames/sec).
To determine the temperature dependent K
Dc
, the Robertson Crack Arrest Temperature
(CAT) test is done, where a crack is forced to run from a low temperature region into a higher
temperature region.
Fig. 8.3 : Instrumented DCB specimen to measure crack propagation speed [39].
Chapter 9
Plastic crack tip zone
The elastic crack tip stresses reach an inﬁnite value, when the distance to the crack tip
decreases to zero. In reality the material will yield before the crack tip is reached and the
elastic solution is no longer valid.
In this chapter the region at the crack tip where yielding occurs is determined. The
inﬂuence of this plastic crack tip zone on the crack behavior is explained. Postyield hardening
is not considered, so the material behaves ideal plastic.
In threedimensional cases, the occurrence of yielding is tested by means of a yield
criterion. Here, we employ the Von Mises and the Tresca criteria. Both are formulated in
terms of principal stresses σ
1
, σ
2
and σ
3
, where the ﬁrst two are in the plane of the plate and
σ
3
is in the normal direction. For plane stress cases σ
3
= 0 and for plane strain situations
Hooke’s law gives σ
3
= ν(σ
1
+σ
2
), with ν as Poisson’s ratio.
9.1 Von Mises and Tresca yield criteria
Von Mises yield criterion is based on the hypothesis that yielding occurs when the speciﬁc
distorsional elastic energy W
d
reaches a critical value W
d
c
. The critical value is determined
from a tensile test and expressed in the yield stress σ
y
. The Von Mises yield criterion can be
expressed in the principal stresses.
Tresca yield criterion is based on the hypothesis that yielding occurs when the maximum
shear stress τ
max
reaches a critical value τ
maxc
. This critical value is again determined in a
tensile test and related to the yield stress σ
y
. The Tresca yield criterion can be expressed in
the diﬀerence of the maximum and the minimum principal stresses.
Von Mises W
d
= W
d
c
(σ
1
−σ
2
)
2
+ (σ
2
−σ
3
)
2
+ (σ
3
−σ
1
)
2
= 2σ
2
y
Tresca τ
max
= τ
maxc
σ
max
−σ
min
= σ
y
83
84
9.2 Principal stresses at the crack tip
For both the Von Mises and the Tresca yield criterion, the principal stresses have to be used.
They are the eigenvalues of the Cauchy stress matrix and for the plane stress case they can
be calculated analytically. For plane strain the third principal stress is not zero but related
to the ﬁrst two principal stresses.
plane stress state σ
zz
= σ
zx
= σ
zy
= 0
σ =
_
_
σ
xx
σ
xy
0
σ
xy
σ
yy
0
0 0 0
_
_
→ det(σ −σI) = 0 →
σ
_
σ
2
−σ(σ
xx
+σ
yy
) + (σ
xx
σ
yy
−σ
2
xy
)
¸
= 0 →
σ
1
=
1
2
(σ
xx
+σ
yy
) +
_
1
4
(σ
xx
−σ
yy
)
2
+σ
2
xy
_
1/2
σ
2
=
1
2
(σ
xx
+σ
yy
) −
_
1
4
(σ
xx
−σ
yy
)
2
+σ
2
xy
_
1/2
σ
3
= 0
plane strain state σ
3
= ν(σ
1
+σ
2
)
Elastic crack tip stress components for Mode I are used to derive the principal crack tip
stresses. The ﬁrst principal stress is always the largest. For plane stress the third principal
stress is zero and the smallest. For plane strain the smallest value may be the second principal
stress, because the value of the third principal stress depends on Poisson’s ratio.
crack tip stresses σ
ij
=
K
I
√
2πr
f
Iij
(θ)
σ
1(+),2(−)
=
K
I
√
2πr
_
cos(
θ
2
)±
_
1
4
_
−2 cos(
θ
2
) sin(
θ
2
) sin(
3θ
2
)
_
2
+
_
sin(
θ
2
) cos(
θ
2
) cos(
3θ
2
)
_
2
_
σ
1
=
K
I
√
2πr
cos(
θ
2
){1 + sin(
θ
2
)} ; σ
2
=
K
I
√
2πr
cos(
θ
2
){1 −sin(
θ
2
)}
σ
3
= 0 or σ
3
=
2νK
I
√
2πr
cos(
θ
2
)
plane stress σ
1
> σ
2
> σ
3
plane strain σ
1
> σ
2
> σ
3
or σ
1
> σ
3
> σ
2
85
0 20 40 60 80 100
0
200
400
600
800
1000
θ
σ
ν = 0.25
0 20 40 60 80 100
0
200
400
600
800
1000
θ
σ
ν = 0.35
0 20 40 60 80 100
0
200
400
600
800
1000
θ
σ
ν = 0.45
0 20 40 60 80 100
0
200
400
600
800
1000
θ
σ
ν = 0.5
Fig. 9.1 : Principal stresses for plane strain as a function of the direction angle. The third
principal stress depends on Poisson’s ratio.
9.3 Von Mises plastic zone
In the Von Mises yield criterion all principal stresses are used. For plane stress and plane
strain the distance r
y
of the elastic/plastic boundary is easily calculated. The derivation uses
the next trigonometric relations :
cos(θ) = cos
2
(
θ
2
) −sin
2
(
θ
2
) = 2 cos
2
(
θ
2
) −1 → cos
2
(
θ
2
) =
1
2
{1 + cos(θ)}
cos
2
(
θ
2
) sin
2
(
θ
2
) =
1
4
sin
2
(θ)
Von Mises yield criterion
(σ
1
−σ
2
)
2
+ (σ
2
−σ
3
)
2
+ (σ
3
−σ
1
)
2
= 2σ
2
y
plane stress σ
3
= 0
(σ
1
−σ
2
)
2
+σ
2
2
+σ
2
1
= 2σ
2
y
K
2
I
2πr
v
cos
2
(
θ
2
)
_
6 sin
2
(
θ
2
) + 2
¸
= 2σ
2
y
r
y
=
K
2
I
2πσ
2
y
cos
2
(
θ
2
)
_
1 + 3 sin
2
(
θ
2
)
¸
=
K
2
I
4πσ
2
y
_
1 + cos(θ) +
3
2
sin
2
(θ)
¸
plane strain σ
3
= ν(σ
1
+σ
2
)
86
(ν
2
−ν + 1)(σ
2
1
+σ
2
2
) + (2ν
2
−2ν −1)σ
1
σ
2
= σ
2
y
K
2
I
2πr
y
cos
2
(
θ
2
)
_
6 sin
2
(
θ
2
) + 2(1 −2ν)
2
¸
= 2σ
2
y
r
y
=
K
2
I
4πσ
2
y
_
(1 −2ν)
2
{1 + cos(θ)} +
3
2
sin
2
(θ)
¸
The elastic/plastic boundaries can be plotted in an xycoordinate system with the crack tip
in the origin and the crack along the line −∞< x < 0; y = 0. It appears that the plane stress
plastic zone is larger than the plane strain plastic zone.
−0.5 0 0.5 1 1.5
−1
−0.5
0
0.5
1
Von Mises plastic zones
pl.stress
pl.strain
Fig. 9.2 : Plastic zones according to the Von Mises yield criterion.
9.4 Tresca plastic zone
Using the Tresca yield criterion implies using the maximum and minimum principal stresses.
For the plane stress case σ
3
= 0 is always the minimum. For the plane strain case, however,
the minimal principal stress can be σ
3
or σ
2
, depending on the coordinate angle θ and on
Poisson’s ratio ν. For plane strain two diﬀerent elastic/plastic boundaries can be calculated.
σ
max
−σ
min
= σ
y
plane stress {σ
max
, σ
min
} = {σ
1
, σ
3
}
87
K
I
_
2πr
y
_
cos(
θ
2
) +
¸
¸
cos(
θ
2
) sin(
θ
2
)
¸
¸
¸
= σ
y
r
y
=
K
2
I
2πσ
2
y
_
cos(
θ
2
) +
¸
¸
cos(
θ
2
) sin(
θ
2
)
¸
¸
¸
2
plane strain I σ
1
> σ
2
> σ
3
→ {σ
max
, σ
min
} = {σ
1
, σ
3
}
r
y
=
K
2
I
2πσ
2
y
_
(1 −2ν) cos(
θ
2
) +
¸
¸
cos(
θ
2
) sin(
θ
2
)
¸
¸
¸
2
plane strain II σ
1
> σ
3
> σ
2
→ {σ
max
, σ
min
} = {σ
1
, σ
2
}
r
y
=
K
2
I
2πσ
2
y
sin
2
(θ)
Plotting the elasticplastic boundary reveals again that the largest plastic zone occurs for the
plane stress case. For the plane strain situation, the plastic region is deﬁned by two curves,
one for σ
1
> σ
2
> σ
3
(smallest) and one for σ
1
> σ
3
> σ
2
(largest).
−0.5 0 0.5 1 1.5
−1
−0.5
0
0.5
1
Tresca plastic zones
pl.stress
pl.strain sig3 = min
pl.strain sig2 = min
Fig. 9.3 : Plastic zones according to the Tresca yield criterion.
9.5 Inﬂuence of the plate thickness
The plastic crack tip zone is considerably diﬀerent for a crack in a thin and a thick plate. In
a thin plate the material can contract freely perpendicular to the plate (3direction), when
88
an inplane load is applied. Over the whole plate thickness a plane stress state exists with
σ
3
= 0. The plane stress plastic crack tip zone develops over the whole plate thickness.
For the thick plate, only the surface layers are free to contract in normal direction and
are thus in a plane stress state. The crack tip material at greater depth from the surface
is prohibited to contract by the surrounding material, which is at a much lower stress and
has no intention at all to contract. The inner crack tip material is thus in a state of plane
strain. In the surface layers we ﬁnd plane stress plastic zones and in the deeper bulk material
the plastic crack tip zone is smaller as it is associated with plane strain. A typical dogbone
plastic zone develops at the crack tip.
Fig. 9.4 : Plastic crack tip zone for a thin and a thick plate.
Critical plate thickness
In a thin plate the relative volume of yielding material is much larger than in the thick plate.
This explains that measured K
Ic
values are higher when determined with thin test specimens.
For thicker specimens, the value is lower and becomes independent of the thickness, when
this is larger than a threshold value B
c
, which is empirically related to K
Ic
and σ
y
.
B
c
>
25
3π
_
K
Ic
σ
y
_
2
> 2.5
_
K
Ic
σ
y
_
2
9.6 Shear planes
Plastic deformation takes place by sliding of crystallographic planes induced by shear stresses.
The planes with the maximum shear stresses can be found at an angle of 45
o
between the
planes with the maximum and minimum principal stresses, which are perpendicular. Em
ploying an etching procedure, the shear planes can be visualized in some cutting planes.
89
Fig. 9.5 : Shear planes at the crack tip: plane stress (left), plane strain (right) [39, 25, 30].
9.7 Plastic constraint factor
When the maximum principal stress is indicated as σ
max
= σ
1
, the other two pricipal stresses
can be written as σ
2
= nσ
1
and σ
3
= mσ
1
. According to the Von Mises yield criterion, σ
max
at yield can be expressed in the initial yield stress σ
y
and a function of the ratios n and m.
This function is the socalled plastic constraint factor pcf. The pcf, which is the ratio of the
maximum principal stress at yield and the yield stress, can also be used as a multiplication
factor of the yield stress to take into account other principal stresses.
_
1
2
{(σ
1
−σ
2
)
2
+ (σ
2
−σ
3
)
2
+ (σ
3
−σ
1
)
2
} =
_
_
1 −n −m+n
2
+m
2
−mn
_
σ
max
= σ
y
pcf =
σ
max
σ
y
=
1
√
1 −n −m+n
2
+m
2
−mn
In a plane stress state with m = 0, we can consider the inﬂuence of the pcf on yielding. For
a range of nratios, the pcf is calulated and the corresponding {σ
1
, σ
2
}values constitute the
plane stress Von Mises yield contour, as expected.
−1 −0.5 0 0.5 1 1.5 2
−1
−0.5
0
0.5
1
1.5
2
σ
1
/σ
y
σ
2
/
σ
y
=
n
σ
1
/
σ
y
Fig. 9.6 : Von Mises plane stress yield contour
90
In the crack tip region the principal stresses are known as a function of the angle θ around
and the distance r to the crack tip. The ratio of σ
2
and σ
1
, n, can be calculated. The ratio
of σ
3
and σ
1
, m, is zero for plane stress and a function of Poisson’s ratio ν for plane strain.
plane stress n =
_
1 −sin(
θ
2
)
¸
/
_
1 + sin(
θ
2
)
¸
; m = 0
plane strain n =
_
1 −sin(
θ
2
)
¸
/
_
1 + sin(
θ
2
)
¸
; m = 2ν/
_
1 + sin(
θ
2
)
¸
In the plane of the crack where θ = 0 we have n = 1. For plane stress we already saw that
m = 0 and for plane strain m = 2ν. The plastic constraint factor can then be determined as
a function of Poisson’s ratio ν. For ν =
1
3
its value is 3, but experiments indicate that it is
smaller. Irwin [36] suggests an over the thickness averaged value pcf =
_
2
√
2 = 1.68.
plane stress pcf = 1 ; plane strain pcf =
1
√
1 −4ν + 4ν
2
9.8 Plastic zone in the crack plane
Considering the plastic zone in the crack plane (θ = 0) for plane stress state, it becomes
clear that the requirement of σ
yy
to be not larger than the yield stress σ
y
, results in loss
of equilibrium. The total internal force in ydirection is decreased, while the external force
remains the same.
9.8.1 Irwin plastic zone correction
In 1958 Irwin [35] has devised a solution to this inconsistency, based on the enlargement of
the plastic zone, such that the total yforce again equals the force associated with the elastic
solution (= shaded area in the ﬁgure).
σ
xx
r
y
a
r
σ
yy
σ
yy
σ
y
σ
xx
r
a
r
p
r
y
σ
y
Fig. 9.7 : Irwin correction of plastic zone in the crack plane.
91
θ = 0 → σ
xx
= σ
yy
=
K
I
√
2πr
yield σ
xx
= σ
yy
= σ
y
→ r
y
=
1
2π
_
K
I
σ
y
_
2
σ
y
r
p
=
_
ry
0
σ
yy
(r) dr =
K
I
√
2π
_
ry
0
r
−
1
2
dr =
2K
I
√
2π
√
r
y
→
r
p
=
2K
I
√
2π
√
r
y
σ
y
→ r
p
=
1
π
_
K
I
σ
y
_
2
= 2 r
y
9.8.2 DugdaleBarenblatt plastic zone correction
In 1960 Dugdale and Barenblatt have published a plastic zone calculation [15], based on the
requirement that in the elastic region outside the plastic zone the stresses must be ﬁnite.
Two load cases are considered to act on the crack with half length a + r
p
, where r
p
is the
plastic zone length to be determined. The ﬁrst load is the Mode I load σ. The second load
is the yield stress σ
y
, applied to the crack faces over the distance r
p
and closing the crack
to its initial half length a. The elastic stresses for x > a + r
p
are known to be a series of
terms of which the ﬁrst term is singular, so leading to inﬁnite stress values. This ﬁrst term
is completely determined by the addition of the two stress intensity factors of the two load
cases. The requirement that this ﬁrst term is zero leads to the value of r
p
.
a
y
x
σ
y
σ
σ
r
p
Fig. 9.8 : DugdaleBarenblatt correction of plastic zone in the crack plane.
load σ K
I
(σ) = σ
_
π(a +r
p
)
load σ
y
K
I
(σ
y
) = 2σ
y
_
a +r
p
π
arccos
_
a
a +r
p
_
92
singular term = 0 → K
I
(σ) = K
I
(σ
y
) →
a
a +r
p
= cos
_
πσ
2σ
y
_
→ r
p
=
πK
2
I
8σ
2
y
9.8.3 Plastic zones in the crack plane
For comparison, the boundary of the elasticplastic transition is considered for θ = 0, so in
the plane of the crack. Poisson’s ratio is ν =
1
3
. The table lists the distance to the crack tip of
the elasticplastic transition. Although Irwin and Dugdale plastic zones were calculated for
plane stress, the plane strain equivalent results from the application of the plastic constraint
factor as a magniﬁcation factor for the yield stress σ
y
.
criterion state r
y
or r
p
r
y
r
p
(K
I
/σ
y
)
2
Von Mises plane stress
1
2π
_
K
I
σ
y
_
2
0.1592
Von Mises plane strain
1
18π
_
K
I
σ
y
_
2
0.0177
Tresca plane stress
1
2π
_
K
I
σ
y
_
2
0.1592
Tresca plane strain σ
1
> σ
2
> σ
3
1
18π
_
K
I
σ
y
_
2
0.0177
Tresca plane strain σ
1
> σ
3
> σ
2
0 0
Irwin plane stress
1
π
_
K
I
σ
y
_
2
0.3183
Irwin plane strain (pcf = 3)
1
π
_
K
I
3σ
y
_
2
0.0354
Dugdale plane stress
π
8
_
K
I
σ
y
_
2
0.3927
Dugdale plane strain (pcf = 3)
π
8
_
K
I
3σ
y
_
2
0.0436
9.9 Small Scale Yielding
The occurrence of a plastic zone at the crack tip rises the question whether the stress intensity
factor K can still be used in a crack growth criterion, as it is associated with a purely linear
elastic solution. The answer is that the elastic stress solution and thus its amplitude K, can
be used when the Kdomain is much larger than the plastic zone. This situation is referred
to as Small Scale Yielding (SSY).
93
The accuracy of the elastic stress ﬁeld in SSY can be enhanced by using an eﬀective
crack length, which takes the plastic zone into account, and is calculated by adding r
y
or r
p
to a. Calculation of K has to be done iteratively, as the plastic zone size depends on K and
vice versa.
a
eff
= a + (r
y
r
p
) ↔ K
I
= β
I
(a
eff
)σ
√
πa
eff
94
Chapter 10
Nonlinear Fracture Mechanics
When Small Scale Yielding cannot be assumed or when the material behavior is intrinsically
nonlinear, concepts from LEFM loose their meaning. For such cases NonLinear Fracture
Mechanics (NLFM) or ElastoPlastic Fracture Mechanics provides new theories and concepts
to analye the behavior of cracks.
10.1 Cracktip opening displacement
In LEFM the displacement of material points in the region around the crack tip can be
calculated. With the crack along the xaxis, the displacement u
y
in ydirection is known
as a function of r (distance) and θ (angle), both for plane stress and for plane strain. The
displacement of points at the upper crack surface results for θ = π and can be expressed in
the coordinate x, by taking r = a − x, where a is the half crack length. The origin of this
xycoordinate system is at the crack center. The crack opening (displacement) (COD) δ is
two times this displacement. It is obvious that the opening at the crack tip (CTOD), δ
t
, is
zero.
u
y
=
σ
√
πa
2µ
_
r
2π
_
sin(
1
2
θ)
_
κ + 1 −2 cos
2
(
1
2
θ)
_¸
displacement in crack plane θ = π; r = a −x
u
y
=
(1 +ν)(κ + 1)
E
σ
2
_
2a(a −x)
Crack Opening Displacement (COD)
δ(x) = 2u
y
(x) =
(1 +ν)(κ + 1)
E
σ
_
2a(a −x)
Crack Tip Opening Displacement (CTOD)
δ
t
= δ(x = a) = 0
Wells (1963) [68] suggested that this CTOD can be used in a crack growth criterion, when
plasticity at the crack tip is taken into account and the actual crack length is replaced by the
eﬀective crack length.
95
96
10.1.1 CTOD by Irwin
The inﬂuence of the crack tip plastic zone can be taken into account by using an eﬀective
crack length a
eff
, which is the actual crack length plus the length of the plastic zone in front
of the crack tip. The correction of Irwin is illustrated in the ﬁgure. Be aware that r
y
is used
here and not the corrected plastic zone r
p
.
σ
xx
r
y
a
r
σ
yy
σ
yy
σ
y
σ
xx
r
a
r
p
r
y
σ
y
Fig. 10.1 : Irwin plastic zone correction.
a
eff
= a +r
y
= a +
1
2π
_
K
I
σ
y
_
2
When the eﬀective crack length of Irwin is used, the crack tip opening can be calculated for
plane stress and for plane strain.
δ(x) =
(1 +ν)(κ + 1)
E
σ
_
2a
eff
(a
eff
−x)
=
(1 +ν)(κ + 1)
E
σ
_
2(a +r
y
)(a +r
y
−x)
δ
t
= δ(x = a) =
(1 +ν)(κ + 1)
E
σ
_
2(a +r
y
)r
y
=
(1 +ν)(κ + 1)
E
σ
_
2ar
y
+ 2r
2
y
≈
(1 +ν)(κ + 1)
E
σ
_
2ar
y
plane stress : δ
t
=
4
π
K
2
I
Eσ
y
=
4
π
G
σ
y
plane strain : δ
t
=
_
1
√
3
_
4(1 −ν
2
)
π
K
2
I
Eσ
y
97
10.1.2 CTOD by Dugdale
In the vision of BarenblattDugdale, the length of the plastic zone follows from the requirement
that the stress at the elasticplastic boundary is not singular.
a
y
x
σ
y
σ
σ
r
p
Fig. 10.2 : Dugdale plastic zone correction.
a
eff
= a +r
p
= a +
π
8
_
K
I
σ
y
_
2
The requirement that the singular term in the stresses at the crack tip must be zero, results
in a crack tip opening which is not zero. It can be calculated for plane stress and for plane
strain.
displacement from requirement ”singular term = 0” : ¯ u
y
(x)
¯ u
y
(x) =
(a +r
p
)σ
y
πE
_
x
a +r
p
ln
_
sin
2
(ˆ γ −γ)
sin
2
(ˆ γ +γ)
_
+ cos(ˆ γ) ln
_
sin(ˆ γ) + sin(γ)
sin(ˆ γ) −sin(γ)
_
2
_
γ = arccos
_
x
a +r
p
_
; ˆ γ =
π
2
σ
σ
y
Crack Tip Opening Displacement
δ
t
= lim
x→a
2¯ u
y
(x) =
8σ
v
a
πE
ln
_
sec
_
π
2
σ
σ
y
__
series expansion & σ ≪σ
y
_
_
_
→
plane stress : δ
t
=
K
2
I
Eσ
y
=
G
σ
y
plane strain : δ
t
=
_
1
2
_
(1 −ν
2
)
K
2
I
Eσ
y
98
10.1.3 CTOD crack growth criterion
In LEFM the CTOD can be related to the energy release rate G and the stress intensity
factor K
I
. In NLFM the CTOD is a measure for the deformation at the crack tip, which can
be compared to a critical value in a crack growth criterion. The critical value, which may
depend on strain rate and/or temperature, has to be measured.
δ
t
= δ
tc
( ˙ ε, T)
10.2 Jintegral
In LEFM, crack growth can be predicted after calculation of the energy release rate G or
the stress intensity factor K. In NLFM, the behavior of a crack can be described with
the Jintegral, which has been introduced by Rice in fracture mechanics in 1968 [54]. The
Jintegral is a vector,
J, with three components in a threedimensional (here Cartesian)
coordinate system. Its value results from an integration along a trajectory Γ. In each point
of this trajectory, the speciﬁc elastic energy must be calculated from stresses and strains.
Also the unit normal vector n in that point, the stress vector
t and the spatial derivatives of
the displacement u play a role.
What is generally called the Jintegral in fracture mechanics, is the ﬁrst component of
the vector
J, so the component in x
1
direction.
Ω
V
x
2
Γ
S
e
1
e
2
x
1
n
t
positive
Fig. 10.3 : Closed curve in a plane with unit normal vector and stress vector.
J
k
=
_
Γ
_
Wn
k
−t
i
∂u
i
∂x
k
_
dΓ ; W =
_
Epq
0
σ
ij
dε
ij
J = J
1
=
_
Γ
_
Wn
1
−t
i
∂u
i
∂x
1
_
dΓ
_
N
m
_
99
10.2.1 Integral along closed curve
The Jintegral vector is deﬁned by integration along a path in a twodimensional plane. When
this path is closed, it can be proved that the resulting value is always zero, under the conditions
that there are no singularities in the area within the closed path. The prove assumes that the
material behavior is hyperelastic and homogeneous, the material is not subjected to volume
loads and the acceleration is zero. It is also necessary to use a linear straindisplacement
deﬁnition, which is only allowed when deformation and rotations are small.
J
k
=
_
Γ
_
Wδ
jk
−σ
ij
u
i,k
_
n
j
dΓ
inside Γ no singularities → Stokes (Gauss in 3D)
_
Ω
_
dW
dε
mn
∂ε
mn
∂x
j
δ
jk
−σ
ij,j
u
i,k
−σ
ij
u
i,kj
_
dΩ
homogeneous hyperelastic σ
mn
=
∂W
∂ε
mn
linear strain ε
mn
=
1
2
(u
m,n
+u
n,m
)
equilibrium equations σ
ij,j
= 0
_
Ω
_
1
2
σ
mn
(u
m,nk
+u
n,mk
) −σ
ij
u
i,kj
_
dΩ =
_
Ω
_
σ
mn
u
m,nk
−σ
ij
u
i,kj
_
dΩ = 0
10.2.2 Path independence
A crack is now introduced along the x
1
axis. The Jintegral – ﬁrst component of the vector
– is calculated along a closed path, which runs around the crack tip and along both crack
surfaces. The integral can be written as a summation of four parts. Obviously the calculated
value is zero.
It is assumed that the crack faces are not loaded (t
i
= 0
i
). On the crack faces, the
unit normal vector is pointing in x
2
direction, so its ﬁrst component n
1
is zero. This result
of the integration along the crack faces is zero, and the Jintegral is the sum of two terms
representing an integration along two lines around the crack tip. When the movement over
the line is anticlockwise, we assign the value a positive value and, when it is clockwise, the
result is negative. With this sign convention, it appears that the value of both integrals has
to be equal. This important result means that the integral over a nonclosed path around the
crack tip, starting at one crack face and ending at the other, always has the same value and
so the Jintegral is path independent.
100
x
1
Ω
n
Γ
−
Γ
B
n
x
2
e
2
e
1
Γ
+
Γ
A
Fig. 10.4 : Closed curve including crack surfaces.
_
Γ
A
f
k
dΓ +
_
Γ
B
f
k
dΓ +
_
Γ
−
f
k
dΓ +
_
Γ
+
f
k
dΓ = 0 →
_
Γ
A
f
1
dΓ +
_
Γ
B
f
1
dΓ +
_
Γ
−
f
1
dΓ +
_
Γ
+
f
1
dΓ = 0
no loading of crack faces : n
1
= 0 ; t
i
= 0 on Γ
+
and Γ
−
_
Γ
A
f
1
dΓ +
_
Γ
B
f
1
dΓ = 0
_
Γ
A
f
1
dΓ = −J
1
A
;
_
Γ
B
f
1
dΓ = J
1
B
−J
1
A
+J
1
B
= 0 → J
1
A
= J
1
B
10.2.3 Relation J ∼ K
In LEFM, the stress and displacement components at the crack tip are known as a function
of the position relative to the crack tip. For multimode loading, they are characterized by
the stress intensity factors K
I
, K
II
and K
III
.
Because the Jintegral is pathindependent, the integration path can be chosen to be
a circle with the crack tip as its center. Integration over this circular path reveals that the
Jintegral is related to the stress intensity factors.
lin. elast. material : W =
1
2
σ
mn
ε
mn
=
1
4
σ
mn
(u
m,n
+u
n,m
)
J
k
=
_
Γ
_
1
4
σ
mn
(u
m.n
+u
n,m
)δ
jk
−σ
ij
u
i,k
_
n
j
dΓ
=
_
Γ
_
1
2
σ
mn
u
m,n
δ
jk
−σ
ij
u
i,k
_
n
j
dΓ
101
Mode I + II + III
σ
ij
=
1
√
2πr
[K
I
f
Iij
+K
II
f
IIij
+K
III
f
IIIij
]
u
i
= u
Ii
+u
IIi
+u
IIIi
substitution and integration over Γ = circle
J
1
=
(κ + 1)(1 +ν)
4E
_
K
2
I
+K
2
II
_
+
(1 +ν)
E
K
2
III
J
2
= −
(κ + 1)(1 +ν)
2E
K
I
K
II
For Mode I loading of the crack, it follows immediately that the Jintegral is equivalent to
the energy release rate G. This means that the Jintegral can be used in the crack growth
criteria of LEFM as a replacement for K and G.
Mode I J
1
= J =
(κ + 1)(1 +ν)
4E
K
2
I
= G
plane stress κ + 1 =
3 −ν
1 +ν
+
1 +ν
1 +ν
=
4
1 +ν
→ J =
1
E
K
2
I
plane strain κ + 1 = 4 −4ν → J =
(1 −ν
2
)
E
K
2
I
Obviously, Jintegral is now also related to the crack tip opening displacement δ
t
.
plane stress Irwin J =
π
4
σ
y
δ
t
Dugbale J = σ
y
δ
t
plane strain Irwin J =
π
4
√
3 σ
y
δ
t
Dugbale J = 2σ
y
δ
t
Based on emperical observations, a better relation is suggested by ASTM, including the
inﬂuence of the specimen dimensions and the ultimate strength σ
u
.
J = mσ
y
δ
t
m = −0.111 + 0.817
a
W
+ 1.36
σ
u
σ
y
10.3 HRR crack tip stresses and strains
Hutchinson, Rice and Rosengren (HRR) derived a solution for the crack tip stress and dis
placement components, when the material behavior is described by the RambergOsgood
relation [31, 55].
102
10.3.1 RambergOsgood material law
Nonlinear material behavior can often be described by an exponential relation between
(equivalent) stress and (equivalent) strain. A much used relation of this type is the Ramberg
Osgood relation. Although it is a general nonlinear material model, we will use it here to
describe postyield behavior. The initial yield stress is σ
y0
and the initial yield strain is
ε
y0
= σ
y0
/E, where E is Youngs modulus, describing the linear elastic behavior. The model
has two parameters, a coeﬃcient α and an exponent n, which is the strain hardening parame
ter. When n = 1 the relation describes linear hardening and with n→∞we have ideal plastic
behavior as is illustrated in the ﬁgure for α = 0.01.
ε
ε
y0
=
σ
σ
y0
+α
_
σ
σ
y0
_
n
0 1 2 3 4 5 6
0
1
2
3
4
5
n = 1
n = 3
n = 5
n = 7
n = 13
ε/ε
y0
σ
/
σ
y
0
Ramberg−Osgood for α = 0.01
Fig. 10.5 : Stressstrain relation according to the RambergOsgood material law.
10.3.2 HRRsolution
The HRRsolution comprises a parameter β, which is directly related to the Jintegral. This
means that the stress and deformation state at the crack tip is completely characterized by
J, just as it is by K in LEFM. The parameter β also depends on a constant I
n
, which value
is determined numerically as a function of the exponent n, both for plane stress and plane
strain.
The solution reveals that the dependency of the distance to the crack tip is determined
by the exponent n. For n = 1 the singularity of the stresses is 1/
√
r, the same as we had in
LEFM. For n > 1 the singularity is smaller.
σ
ij
= σ
y0
β r
−
1
n+1
˜ σ
ij
(θ) ; u
i
= αε
y0
β
n
r
1
n+1
˜ u
i
(θ)
with : β =
_
J
ασ
y0
ε
y0
I
n
_
1
n+1
(I
n
from num. anal.)
103
5
2.5
5 10 15
0
plane strain
plane stress
n
I
n
Fig. 10.6 : Value of the variable I
n
for plane strain and plane stress.
10.3.3 Jintegral crack growth criterion
We saw that the Jintegral can replace the energy release rate in LEFM and is related to the
stress intensity factor. In NLFM, where the material behavior is described by the general
RambergOsgood relation, the Jintegral characterizes the stress at the crack tip. It is thus
obvious that it can be used in a crack growth criterion. Calculation of its value is easily done,
due to the fact that the integration path can be chosen arbitrarily. Critical values have to be
measured according to normalized experiments.
J = J
c
104
Chapter 11
Numerical fracture mechanics
In the crack growth criteria of LEFM and NLFM, a crack growth parameter has to be calcu
lated and compared to a critical value, which is known from experiments. Crack parameters,
G, K, δ
t
and J, can be calculated analytically, but this is only possible with great mathemat
ical eﬀort and even then for rather simple cases. When such solutions exist, they can often
be found in literature. Using numerical methods, it is possible to calculate their values rather
straightforwardly, also for practical problems.
Finite element programs are widely available and can be used for the numerical calcula
tion of crack growth parameters. Although special procedures are developed and implemented,
the use of standard elements is possible after a minor adaptation of the element mesh. Cal
culation of the Jintegral is often implemented as a standard option.
With the ﬁnite element method it is also possible to simulate crack propagation. Several
techniques are used and new ones are still under development.
Besides the ﬁnite element method, the boundary element method is used in numerical
fracture mechanics. Commercial programs for BEM are available although not so widespread
as for FEM.
11.1 Quadratic elements
For twodimensional ﬁnite element analyses, the 8node plane stress or plane strain element
gives accurate results for most mechanical problems. The displacement of an internal element
point is interpolated quadratically between the nodal displacements. Although initially the
element edges are straight lines, after deformation they may become of parabolic shape.
Interpolation or shape functions are used for this purpose, which are a function of two
local socalled isoparametric coordinates ξ
1
and ξ
2
, which have values between 1 and +1.
The ξ
1
, ξ
2
coordinate system is not orthonormal, because the axes are not perpendicular.
Strains, which are derivatives of the displacements, are more or less linear in the local
coordinates and this is also the case for the stresses, which are proportional to the strains. It
has to be said that these statements are not valid when nonlinear straindisplacement and
stressstrain relations are used.
105
106
ξ
1
4
7
3
6
ξ
1
ξ
2
2 5
1
8
ξ
2
8
4
7
3
6
2 5
1
8
4
7
3
6
5
2
1
Fig. 11.1 : Quadratic elements with 8 nodal points.
11.2 Crack tip mesh
The ﬁgure shows part of the element mesh around the crack tip. Where the crack has opened,
the nodes of elements above and below the crack may coincide, but must not be connected.
When the material is assumed to be linearly elastic, theory says that the stresses at
the crack tip are proportional to 1/
√
r, with r the distance to the crack tip. The very
steep stress gradient at the crack tip can only be described with a lot of small elements.
However, it is found that reﬁning the mesh at the crack tip will not lead to convergence,
i.e. further reﬁnement will always render diﬀerent results. The reason for this unacceptable
mesh dependency is the singularity in the stress ﬁeld, which is not described by the element
interpolation functions.
Fig. 11.2 : Crack tip mesh.
107
11.3 Special elements
Solutions to circumvent the shortcomings of the standard 8node element to describe the
crack tip stresses well, are still being sought and found.
One solution is to change the interpolation function of the element in such a way that
the crack tip ﬁeld is included. This results in a socalled enriched element, which is only
usable at the crack tip.
Another solution may be found in application of a modiﬁed variational principle as the
basis for the ﬁnite element equations. The resulting element is a hybrid element, where stresses
are unknown nodal variables.
Using enriched or hybrid elements implies implementation of these elements and some
times new solution techniques in the FEM program, which may be very unfeasable, when a
commercial program is used. However, there is a possibility to calculate the crack tip stress
ﬁeld very accurately with standard 8node elements, which are slightly adapted into socalled
quarter point elements.
11.4 Quarter point elements
Quarter point elements are devised by Barsoum in 1976 [2] as a simple means to describe
the stress singularity at the crack tip. They are standard 8node quadrilateral or 6node
triangular elements, where two midside points are repositioned towards one corner node,
such that they divide there side in the ratio 1:3. They are used at the crack tip, which is
located in the node towards which the midside nodes are moved. It can be proved that they
describe the 1/
√
rstress ﬁeld singularity at the crack tip accurately.
Quarter point elements are also referred to as Distorted Quadratic Quadrilateral or
Triangle. Another suitable solution is to use a Collapsed Quadratic Quadrilateral, where
three points of the 8node quadrilateral element are situated at the same location, but not
connected.
A linear 4node quadrilateral element can be collapsed into a triangle and it appears to
describe a 1/r singularity very accurately.
Quarter point elements can be made very easily at the crack tip by replacing some mid
side nodes of the element mesh. The big advantage is that they allow the use of standard
ﬁnite element packages.
1 5 2
6
3 7 4
8
p 3p
1
2
5
3
6
4
1
5 2
6
3
7
3p p
4
8
2
1
4
3
Fig. 11.3 : Quarter point and collapsed elements.
108
The crack tip mesh is often made as a crack tip rozet, with quarter point elements at the
crack tip. Lim advised in 1993 [43], to use transition elements between the quarter point
elements and the regular mid point elements for better results.
Fig. 11.4 : Crack tip rozet.
11.4.1 Onedimensional case
For a onedimensional case, it will be proved that the quarter point technique indeed renders
the 1/
√
r singularity. The ﬁgure shows a line element with three nodes and a local coordinate
−1 ≤ ξ ≤ 1. The global coordinate x has its origin in node 1, where the crack tip can be
seen.
The global position of an element point can be interpolated between the position of
the three nodes {x
1
, x
2
, x
3
}, where x
1
= 0 and x
2
= L, the length of the element. The
displacement u in xdirection is interpolated in the same way between the nodal displacements
u
1
, u
2
and u
3
. The strain of the element is calculated as the derivative of u w.r.t. x, which
can be calculated, using the interpolations.
1
x
2 3
ξ = 0 ξ = −1 ξ = 1
Fig. 11.5 : Onedimensional threepoint element at the crack tip.
x =
1
2
ξ(ξ −1)x
1
+
1
2
ξ(ξ + 1)x
2
−(ξ
2
−1)x
3
=
1
2
ξ(ξ + 1)L −(ξ
2
−1)x
3
109
u =
1
2
ξ(ξ −1)u
1
+
1
2
ξ(ξ + 1)u
2
−(ξ
2
−1)u
3
du
dξ
= (ξ −
1
2
)u
1
+ (ξ +
1
2
)u
2
−2ξu
3
→
du
dx
=
du
dξ
dξ
dx
=
du
dξ
/
dx
dξ
Mid point element
When node 3 is located in the center of the element (x
3
=
1
2
L), the strain in node 1 (at the
crack tip) can be evaluated and appears to be a linear function of the local coordinate ξ. The
strain at node 1 (the crack tip) is constant.
ξ = −1
1
x
3 2
ξ = 0 ξ = 1
Fig. 11.6 : Onedimensional mid point element at the crack tip.
x =
1
2
ξ(ξ + 1)L −(ξ
2
−1)
1
2
L =
1
2
(ξ + 1)L ⇒
dx
dξ
=
1
2
L
du
dx
=
du
dξ
1
2
L
→
du
dx
¸
¸
¸
¸ x=0
ξ=−1
=
_
2
L
_
__
−
3
2
_
u
1
+
_
1
2
_
u
2
+ 2u
3
_
Quarter point element
When node 3 is positioned at the quarter of the element length (x
3
=
1
4
L), the strain appears
to be proportional to 1/
√
x and becomes inﬁnite in node 1.
x
3 2
ξ = −1 ξ = 1 ξ = 0
1
Fig. 11.7 : Onedimensional quarter point element at the crack tip.
x =
1
2
ξ(ξ + 1)L −(ξ
2
−1)
1
4
L =
1
4
(ξ + 1)
2
L → ξ + 1 =
_
4x
L
⇒
dx
dξ
=
1
2
(ξ + 1)L =
√
xL
du
dx
=
du
dξ
√
xL
→
du
dx
¸
¸
¸
¸ x=0
ξ=−1
= ∞ → singularity
1
√
x
110
11.5 Virtual crack extension method (VCEM)
One of the variables which can be calculated with a standard ﬁnite element package is the
energy release rate G. It is deﬁned by diﬀerentiation of internal and external work w.r.t. the
crack length.
The energy release rate can be calculated following two analyses, with a small diﬀerence
in crack length ∆a. The ratio of the diﬀerence in elastic energy and ∆a is a good approxima
tion for G, when ∆a is indeed very small and when ”ﬁxed grips” is used, i.e. when boundary
displacements are prescribed.
This procedure is called the Virtual Extension Method or Stiﬀness Derivation Method.
a +∆a
a
u u
I II
Fig. 11.8 : Virtual Crack Extension Method.
ﬁxed grips →
dU
e
da
= 0 ⇒ G = −
1
B
dU
i
da
≈ −
1
B
U
i
(a +∆a) −U
i
(a)
∆a
stiﬀness matrix variation
Instead of analyzing the whole structure two times with small diﬀerence in crack length, the
energy release rate can be calculated directly from the stiﬀness matrix. Only a small part of
the mesh around the crack tip is considered and its contribution to the global stiﬀness matrix
C is calculated for crack length a and for crack length a + ∆a, where the outer boundary
of the crack tip mesh is ﬁxed. The change in the stiﬀness matrix can be used directly to
evaluate G. Because only a small region around the crack tip is considered, it is important
to use quarter point elements at the crack tip.
Fig. 11.9 : VCEM in crack tip mesh.
111
BG = −
dU
i
da
= −
1
2
u
˜
T
∆C
∆a
u
˜
with ∆C = C(a +∆a) −C(a)
11.6 Stress intensity factor
Calculation of the stress intensity factor is of course possible trough its relation with the
energy release rate.
plane stress K
2
I
= EG
I
; K
2
II
= EG
II
plane strain K
2
I
=
E
1 −ν
2
G
I
; K
2
II
=
E
1 −ν
2
G
II
Calculation of the stress intensity factor is also possible through its deﬁnition, which relates it
to the stress ﬁeld. Stress components have to be calculated accurately upon which K results
from extrapolation to the crack tip. Because stresses are calculated in integration points and
the crack tip is located in an element node, this calculation is not exactly according to the
deﬁnition of K. It always remains to be investigated how the interpolation has to be done to
get the best results.
K
I
= lim
r→0
_
√
2πr σ
22

θ=0
_
; K
II
= lim
r→0
_
√
2πr σ
12

θ=0
_
θ
r
K
p1
K
p2
K
p3
K
p4
K
r
1
r
2
r
3
r
4
r
p
1
p
2
p
3
p
4
Fig. 11.10 : Calculation of K by extrapolation.
Stresses in element integration points are calculated by diﬀerentiation, which makes them less
accurate then the calculated nodal displacements. It seems logical therefore to calculate the
stress intensity factor not from the stress ﬁeld but from the displacement ﬁeld.
u
y
=
4(1 −ν
2
)
E
_
r
2π
K
I
g
ij
(θ) → K
I
= lim
r→0
_
E
4(1 −ν
2
)
_
2π
r
u
y
(θ = 0)
_
112
11.7 Jintegral
The Jintegral can be calculated according to its deﬁnition, i.e. by integration along a path
around the crack tip. Because stresses and strains are known in integration points, the
integration path must preferably pass through these points. In a general crack tip mesh such
a path would be diﬃcult to deﬁne during postprocessing of the results.
Fig. 11.11 : Jintegral path in crack tip rozet.
In a regular crack tip mesh, an integration path could be deﬁned as indicated in the ﬁgure.
It consists of lines in coordinate directions. When symmetry can be considered, only half of
the path has to be taken into account. When such a path is taken at a certain distance from
the crack tip, there is no need to use quarter point crack tip elements.
Fig. 11.12 : Jintegral path through integration points.
11.7.1 Domain integration
In commercial element packages, the Jintegral over a curve Γ is transformed to an integral
over the domain Ω, enclosed by Γ. Because the path Γ is not a closed path, Stokes theorem
cannot be applied directly. In a paper by Li et.al. published in 1985, it is shown that the
transformation is possible when a shift function q is introduced, which is 0 at the outer
boundary of Ω and 1 at the inner boundary.
113
x
1
Ω
n
Γ
−
Γ
B
n
x
2
e
2
e
1
Γ
+
Γ
A
Ω
q = 0
q = 1
Fig. 11.13 : Jintegral domain.
J =
_
Ω
∂q
∂x
j
_
σ
ij
∂u
i
∂x
1
−Wδ
1j
_
dΩ
11.7.2 De Lorenzi Jintegral : VCE technique
In the Marc/Mentat package a general Jintegral is calculated, which includes phenomena
like pressure on the crack faces p
i
, acceleration eﬀects and initial strains ε
o
ij
. It is calculated as
a domain integral, where the shift function q is deﬁned as the ratio of a displacement ∆x
1
of
nodes within the domain Ω and the elongation ∆a of the crack. The user has to indicate which
ring of elements around the crack tip must be moved when the crack is virtually extended.
In one analysis several options can be chosen.
Calculating such Jintegral values is a default option in major FEA codes. It does not
require intensive mesh reﬁnement near the crack tip or the use of quarter point elements.
Fig. 11.14 : Jintegral domain with various element rings.
114
J =
_
Ω
∂q
∂x
j
_
σ
i1
∂u
i
∂x
1
−Wδ
1j
_
dΩ −
_
Γs
qp
i
∂u
i
∂x
1
dΓ −
_
Ω
q(ρq
i
−ρ¨ u
i
)
∂u
i
∂x
1
dΩ +
_
Ω
qσ
ij
∂ε
o
ij
∂x
1
dΩ
11.8 Crack growth simulation
Numerical procedures can not only be used to evaluate crack growth criteria. The actual crack
propagation can be simulated. Various approaches exist and they are improved in ongoing
research.
11.8.1 Node release
One of the techniques to simulate crack propagation is the ”node release” method. When
a crack growth criterion indicates that a crack will growth, use of a crack growth direction
criterion like MTS or SED, reveals in what direction the crack tip wants to move. In the node
release approach, a crack can only move along element edges and the movement is realized
by splitting one or more nodes in two, which are not connected anymore. The resulting crack
path is strongly depending on the mesh geometry. Application is not so easy to accomplish,
because the propagation implies that the element mesh topology has to change.
By introducing springs and/or dashpots between the released nodes, ductile crack prop
agation can be simulated.
Fig. 11.15 : Node release method for crack growth simulation.
11.8.2 Moving Crack Tip Mesh
Because of the high stress gradients at the crack tip, small elements are required there, while
elements at a some distance to the crack tip can be considerably larger. It is for this reason
that a socalled crack tip mesh (CTM) is often used. When crack propagation is simulated,
115
this CTM can follow the crack tip. The example shows the crack propagation in the interface
between ﬁber and matrix for a plane strain simulation in the transversal plane, where the
loading is in the vertical direction. The crack opening is shown with enlarged displacements.
It was possible to simulate the crack propagation, which initially was in the ﬁbermatrix
interface but from some point moved into the matrix material. The CTM was moved by hand
for each crack advancement, which was a very laborious operation.
Fig. 11.16 : Crack tip mesh moves along ﬁbermatrix interface [42].
11.8.3 Element splitting
Instead of advancing a crack along element edges, propagation through elements is also an
option. It implies the subdivision of an element in two or more smaller elements, which
changes the mesh topology. Results are more accurate than those of the node release method.
Mesh reﬁnement can be realized in the region of the crack tip.
116
Fig. 11.17 : Crack propagation by element splitting technique [46].
11.8.4 Smeared crack approach
The smeared crack approach can be seen as an intermediate between fracture mechanics and
continuum damage mechanics. Attention is not focused on one large crack, but the material
damage is modeled with small cracks in element integration points. These cracks may initiate
and lengthen according to certain loading criteria. The cracks are not really modeled, but
simulated with anisotropic material behavior, where in one direction the material stiﬀness is
reduced.
117
e
1
e
2
n
2
σ
1
σ
2
n
1
n
1
n
2
e
1
e
2
Fig. 11.18 : Smeared crack approach for crack propagation.
118
Chapter 12
Fatigue
Many structures are subjected to cyclic loading, with an amplitude so low that failure due to
yielding or crack growth, is never to be expected. However, experience tells a diﬀerent story.
After a large number of load reversals, structural parts fail unexpectedly, often leading to
dramatic and overall collapse. This very dangerous phenomenon is the result of fatigue crack
growth.
Fatigue failure became an issue when carriages and trains with iron or steel axles began
traveling frequently over long distances in the mid 19th century. Axles broke mostly at
diameterjumps and this was explained ﬁrst as the result of changes in the material’s structure.
W¨ohler was among the ﬁrst to investigate these phenomena systematically, all of course
experimentally. Over many years, a tremendous amount of data was gathered, but theoretical
explanation had to wait until Griﬃth saw the importance of cracks for the failure of materials.
Cyclic loading of structures does not only occur as the result of deliberately applied
load reversals. Vibrations (machine(s) (parts)), repeated pressurization en depressurization
(airplanes), thermal cycling (switching of electronic devices), random forces (ships, vehicles,
planes) occur always and everywhere.
12.1 Crack surface
When fatigue failure has occurred, the crack surface shows two distinct markings on two
diﬀerent scales.
At a macroscopic scale, socalled clam shell markings also called beach marks can be
seen. They are the result of irregularities in the growth of the fatigue crack, due to changes
in loading conditions.
When the crack surface is observed with a microscope, the crack displacement during each
individual load cycle is seen as a pattern of striations. During one cycle, plastic deformation
at the crack tip results in sequential blunting and sharpening of the crack tip. This crack
growth is very regular.
119
120
Fig. 12.1 : Clam shell fatigue crack surface and striations [29].
12.2 Experiments
Experiments can be done on fullscale structures subjected to reallife loading. This, however,
is often not feasible, due to costs and the extremely long times involved before the detection
of fatigue damage. Experiments are therefore mostly done on laboratory scale, using small
test specimens subjected to high loads varying at high frequencies. Results must be carefully
evaluated, because size eﬀects and upscaling of load amplitude and frequencies may lead to
inconsistencies with reallife applications.
12.3 Fatigue load
When fatigue crack growth is studied experimentally, the loading is ﬁrst taken to be harmonic
and is characterized in terms of stresses or strains. For stress controlled loading, the stress
range ∆σ is the diﬀerence between the maximum and the minimum stress during the cycle.
The stress amplitude σ
a
is half the stress variation. Average stress σ
m
and stress ratio R
σ
are used to specify the loading. Also the amplitude ratio σ
a
/σ
m
is sometimes used.
In strain controlled experiments, all these variables obviously apply to strains.
N
σ
m
t
σ
max
σ
min
i + 1 i
0
0
σ
Fig. 12.2 : Harmonic stress.
121
∆σ = σ
max
−σ
min
; σ
a
=
1
2
∆σ
σ
m
=
1
2
(σ
max
+σ
min
) ; R
σ
= σ
min
/σ
max
;
σ
a
σ
m
=
1 −R
1 +R
12.3.1 Fatigue limit
Experiments reveal that some materials – e.g. mild steel, low strength alloys, some Ti/Al/Mg
alloys – never fail under cyclic loading, when the stress in each cycle is below a certain
threshold σ
th
, also referred to as fatigue limit. Other materials – high strength steels, most
nonferro alloys, some austenitic steels, some Ti/Al/Mg alloys – do not have such a fatigue
limit: they will always fail no matter how low the stress, although it may take a very large
number of load cycles.
σ
N
σ
th
Fig. 12.3 : Harmonic stress and fatigue limit.
12.3.2 (SN)curve
The experimental data are represented as a (SN) or W¨ohler curve, where the maximum
stress S = σ
max
is indicated along the vertical axis and the logarithm of the number of cycles
to failure or fatigue life N
f
along the horizontal axis. Determination and representation of
these data is prescribed in normalization sheet B.S. 3518 part I 1984.
These curves are mostly recorded for zero average stress. For materials with a fatigue
limit, the (SN)curve will advance towards a horizontal asymptote at the level σ = σ
th
. When
a fatigue limit does not exist, the fatigue strength or endurance limit is deﬁned as the stress
variation for failure after a speciﬁed high – typically 10
6
– number of cycles. The ratio of
fatigue limit or endurance limit to tensile strength is referred to as endurance ratio.
122
0
0
S
log(N
f
)
σ
th
Fig. 12.4 : SNcurve with fatigue limit.
Fatigue data may also be presented in a (S
a
N)curve, where the vertical axis indicates the
stress amplitude : S
a
= σ
a
=
1
2
∆σ.
For some real materials the (SN)curves are shown in the ﬁgure.
10
4
10
5
10
6
10
7
10
8
10
9
100
150
200
250
300
350
400
450
500
550
600
steelT1
Al2024T4
Mgalloy
steel1020
N
f
σ
m
a
x
[
M
P
a
]
Fig. 12.5 : SNcurves for various materials.
12.3.3 Inﬂuence of average stress
When the average stress is positive, a crack will grow faster and the curve will be located at
lower stress values, indicating that the fatigue life N
f
is reached earlier for the same stress
amplitude σ
a
. It can also be stated that for the same N
f
, σ
a
is reduced to σ
∗
a
.
123
0
0
log(N
f
)
σ
m
σ
th
σ
a
Fig. 12.6 : Inﬂuence of the average stress on the S
a
Ncurve.
Corrections have been proposed to take into account the decrease in σ
a
due to σ
m
= 0 and to
relate it to the tensile strength σ
u
or the initial yield stress σ
y0
. Much used are the emperical
formulas of Gerber, Goodman and Soderberg.
Gerber (1874)
σ
∗
a
σ
a
= 1 −
_
σ
m
σ
u
_
2
Goodman (1899)
σ
∗
a
σ
a
= 1 −
σ
m
σ
u
Soderberg (1939)
σ
∗
a
σ
a
= 1 −
σ
m
σ
y0
12.3.4 (PSN)curve
Variations in experimental fatigue data can be represented in a (PSN)curve, where proba
bility (P) of failure is incorporated. The center curve represents the stress for which 50 % of
all specimens will fail after the associated N
f
. When the upper curve is reached, 99 % of all
specimens will have failed. The lower curve represents the weakest specimens. For the given
stress, 1 % of all specimens fail for N
f
on this curve.
10
4
10
5
10
6
10
7
10
8
100
150
200
250
300
350
400
450
500
550
50% prob.failure
90% prob.failure
10% prob.failure
N
f
σ
m
a
x
[
M
P
a
]
Fig. 12.7 : PSNcurves.
124
12.3.5 High/low cycle fatigue
For low stresses and thus long fatigue life (typically N
f
> 50000), we are in the High Cycle
Fatigue (HCF) range. Because stresses are very low, the plastic crack tip zone is small and
concepts from LEFM (+ SSY) can be applied. We can characterize the crack tip stress ﬁeld
with stress intensity factors.
For high stresses and thus short fatigue life, we are in the Low Cycle Fatigue (LCF) range.
Crack tip stresses are so high that NLFM has to be applied to study the crack behavior.
For high cycle fatigue the (SN)curve can be ﬁtted with the Basquin relation. For low
cycle fatigue the MansonCoﬃn relation is used for the ﬁt.
0
S
a
0
log(N
f
)
4 5
LCF HCF
σ
m
= 0
Fig. 12.8 : Low and High Cycle Fatigue regimes.
12.3.6 Basquin relation
The Basquin relation relates the stress amplitude to the number of load reversals, which is
twice the number of cycles. The relation has two parameters: the fatigue strength coeﬃcient
σ
′
f
, and the fatigue strength exponent or Basquin exponent b.
1
2
∆σ = σ
a
= σ
′
f
(2N
f
)
b
→ ∆σN
−b
f
= constant
log
_
∆σ
2
_
log(2N
f
)
Fig. 12.9 : Basquin ﬁt of SNcurve.
125
12.3.7 MansonCoﬃn relation
The MansonCoﬃn relation relates the plastic strain amplitude to the number of load re
versals, which is twice the number of cycles. The relation has two parameters: the fatigue
ductility coeﬃcient ε
′
f
, and the fatigue ductility exponent or MansonCoﬃn exponent c.
1
2
∆ε
p
= ε
′
f
(2N
f
)
c
→ ∆ε
p
N
−c
f
= constant
log
_
∆ε
p
2
_
log(2N
f
)
Fig. 12.10 : MansonCoﬃn ﬁt of SNcurve.
12.3.8 Total strainlife curve
Basquin and MansonCoﬃn relations can be combined. To that purpose, the stress amplitude
in the Basquin relation has to be replaced by a strain amplitude, which is easily accomplished,
because the low stresses assure linear elastic behavior. The total strain amplitude is the
addition of the elastic and plastic strain amplitudes, which is a valid assumption for small
deformations.
The resulting total strainlife curve can be plotted, using logarithmic axes.
log(
∆ε
2
)
log(N
f
)
Fig. 12.11 : Total strainlife curve.
∆ε
2
=
∆ε
e
2
+
∆ε
p
2
=
1
E
σ
′
f
(2N
f
)
b
+ε
′
f
(2N
f
)
c
126
12.4 Inﬂuence factors
Fatigue life of a test specimen is determined by the external load amplitude and frequency, as
is experimentally represented in an (SN)curve. For real structures, the fatigue life is highly
determined by macroscopic shape and dimension, and by microscopic conﬁguration (material
microsructure). Manufacturing procedures are also very inﬂuential. Environmental inﬂuences
can have a high impact on the initiation and propagation of fatigue damage as well.
12.4.1 Load spectrum
Fatigue life is experimentally represented in an (SN)curve. Until so far we considered tensile
loading where S can be the maximum stress or the stress amplitude. The inﬂuence of the
average stress σ
m
is obvious and has been discussed.
Fatigue data are independent of the loading frequency within certain limits. Too low or
too high frequencies will result in diﬀerent fatigue life.
Shear loading experiments result in diﬀerent (SN)curves than the tensile equivalents.
Multiaxial loading results in lower fatigue life that uniaxial loading.
12.4.2 Stress concentrations
The shape of structural parts, combined with the external loading, almost always leads to
inhomogeneous deformation and nonuniform stress ﬁelds. In certain situations local stresses
can become much higher than the average or nominal stress and such stress concentrations
are mostly found at notches. The stress concentration factor K
t
is the ratio between the
maximum stress and the nominal stress and is a function of the notch radius ρ.
When the specimen is subjected to an harmonic load, the fatigue or endurance limit
is theoretically reduced with 1/K
t
. Experiments have shown, however, that the fatigue life
is reduced less severely than this: it is reduced by a factor 1/K
f
, where K
f
is the fatigue
strength reduction factor, which is smaller than K
t
.
The notch sensitivity factor q relates K
f
to K
t
and is found to be a function of the notch
radius and a material parameter. Relations for q have been proposed by Peterson [53] and
Neuber [50].
ρ
Fig. 12.12 : Stress concentration at axis diameter reduction.
∆σ
th
(notched) =
1
K
f
∆σ
th
(unnotched) ; 1 < K
f
< K
t
127
K
f
= 1 +q(K
t
−1)
Peterson : q =
1
1 +
a
ρ
with a = material parameter
Neuber : q =
1
1 +
_
b
ρ
with b = grain size parameter
12.4.3 Stress gradients
The size of a structural part, e.g. an axis subjected to bending, determines the gradients of
stresses at equal external load. Larger specimens show lower gradients than smaller ones,
when subjected to the same external load. When the gradient is lower, more material is
subjected to a raised stress level than in the case of a higher gradient. This results in a
decrease of fatigue life with specimen size. This size eﬀect often necessitates testing in full
scale experiments, especially for critical applications like aviation.
Fig. 12.13 : Stress gradients in axes of diﬀerent diameter.
12.4.4 Material properties
The ease of fatigue crack initiation and growth is obviously inﬂuenced by the microstructure
of the material. At low temperatures grain boundaries act as a barrier for crack propagation,
so fatigue limit is higher for small grain microstructures. At high temperatures, the grain
boundaries are weakened by diﬀusion of atoms (creep), so fatigue limit is lower for small grains.
The texture (grain shape) also inﬂuences fatigue life and leads to anisotropy. Inhomogeneities
– lattice imperfection, voids, particles, ﬁbers – may have a negative (stress concentration →
initiation of cracks) or positive (crack arrest, dissipation) inﬂuence on fatigue life.
128
12.4.5 Surface quality
Material microstructure determines the growth of a fatigue crack to a high extent. At the
external surface of the specimen, crystal lattices may be squeezed out by cyclic plastic defor
mation. The resulting extrusions (and intrusions) act as a notch, where stress concentrations
will easily lead to the initiation of a fatigue crack. They show up at about 0.1 N
f
.
The surface roughness can be set during manufacturing (or specimen preparation) and
will highly inﬂuence the surface crack initiation and fatigue life. Surface treatments, like shot
peening and carbonizing, can be applied to induce compressive stresses, which will prohibit
crack initiation and growth. A coating can be applied to eliminate deteriorating environmen
tal eﬀects. Increasing the yield stress by applying alloying materials, will result in a higher
resistance against lattice slip.
Besides the external surface, internal surfaces at bulk defects, are also inﬂuencing fatigue
life.
10µm
Fig. 12.14 : In and extrusion of lattices at the surface.
12.4.6 Environment
Environmental inﬂuences on fatigue life are mostly due to temperature eﬀects and chemical
attack.
Low temperatures lead to embrittlement, due to the prevention of plastic deformation.
High temperature (T > 0.5T
m
) facilitates the creep deformation due to diﬀusion of atoms to
and from grain boundaries, dislocation movement and migration of vacancies. Weakening of
the grain boundaries leads to loss of intergranular adhesion. Fatigue cracks will travel more
easily between grains or cross grain boundaries. This phenomenon is referred to as creep
fatigue.
Grain boundaries may also be weakened by chemical inﬂuence. Hydrogen and other
gasses can diﬀuse to the grain boundaries and diminish bond strength. Protective surface
ﬁlms may be cracked. This process is called corrosion fatigue.
12.5 Crack growth
Detailed experiments have shown that the crack length a is an exponential function of the
number of cycles N. This means that crack growth is very slow until the ﬁnal stage in the
129
fatigue life, where a relative short number of cycles will result in fast crack growth leading
to failure. The initial fatigue crack length a
i
seems to be a very important parameter for the
fatigue life N
f
.
For an initially undamaged material, it takes N
i
cycles to initiate a crack by dislocation
movement and void coalescence. At this fatigue crack initiation life the initial crack has been
formed, but in most cases it is so small that it cannot be detected. In this stage I, crack
growth is provoked by shear stresses and involves slip on a single crystallographic slip plane.
The crack propagation rate is very low, typically < 0.25 nm/cycle.
After N
i
cycles, in stage II of crack growth, crack propagation is faster, typically µm’s
per cycle. The crack growth is provoked by tensile stresses and involves plastic slip on multiple
slip planes at the crack tip, resulting in striations. After a large number of cycles the crack
reaches a length a
1
, which can be detected by nondestructive techniques. The crack growth
is now much faster and after the fatigue life N
f
its length is a
f
and after a few cycles a
c
the
critical crack length is reached and failure occurs. For higher loading amplitudes, the crack
growth will be faster.
After N cycles, the cycles to go until failure at N
f
, is indicated as N
r
. The restlife is
the ratio of N
r
and N
f
.
N
σ
a
c
a
i
N
i
N
f
a
f
a
c
a
1
a II I III
Fig. 12.15 : Crack length increase with number of cycles.
N
r
N
f
= 1 −
N
N
f
12.5.1 Crack growth models
To predict the fatigue life of structures, crack growth models have been proposed, which relate
the crack grow rate
da
dN
to load amplitude or maximum load, which can be expressed in the
stress intensity factor K, because we assume to be in the high cycle fatigue regime, where
stresses are low.
Microstructural models relate the crack grow rate to microstructural parameters, such
as the spacing between the striations. Bates and Clark [3] found that for certain materials
this spacing is about 6 times (∆K/E)
2
.
Phenomenological models are formulated and ﬁtted by close inspection of experimental
crack growth data. All these relations describe
da
dN
to be a function of the stress σ and the
130
crack length a. This is generally written as a relation between
da
dN
and ∆K, where scaling
with E improves the generality of the ﬁtted results. However, in LEFM,
da
dN
is mostly related
to ∆K by the socalled Paris law, using two parameters : a coeﬃcient and an exponent,
da
dN
= C(∆K)
m
12.5.2 Paris law
Paris & Erdogan published the crack grow law, which has become known as the Paris law in
1963 and it is still widely used [51]. On a doublelogarithmic axes, the Paris law is represented
by a straight line. The two parameters, C and m, can be ﬁtted easily, when two data points
are known, as is shown in the example below. Their values depend on material, geometry,
load and loading frequency.
6
7
8
9
5
1
2
3
4
1
[MPa
√
m]
log
_
da
dN
_
[mm/c]
log C = −8.7
log(∆K)
2 0 4 3
Fig. 12.16 : Paris law parameter ﬁt.
da
dN
= C(∆K)
m
→ log
_
da
dN
_
= log(C) +m log(∆K)
log(∆K) = 0 → log(C) = log
_
da
dN
_
= −8.7 → C = 2 ×10
−9
[mm]
[MPa
√
m]
m
m =
(−2) −(−4)
(2) −(1.5)
= 4
Limits of Paris law
For low and high values of ∆K, Paris law will not describe the crack growth rate accurately
any more. For ∆K ≈ K
th
, the lower limit, a crack grows extremely slowly, hampered by the
roughness of the crack faces. For still smaller values of ∆K, the crack growth is extremely
small but not completely zero. For high values of ∆K, crack growth is much faster than
predicted by the Paris law.
131
rapid crack growth
∆K
th
∆K
c
log(∆K)
R
log(
da
dN
)
slow crack growth
power law growth
Fig. 12.17 : Limits of Paris law.
Paris law parameters
For some materials values of Paris law parameters are listed.
material ∆K
th
[MNm
−3/2
] m[] C×10
−11
[!]
mild steel 3.2  6.6 3.3 0.24
structural steel 2.0  5.0 3.85  4.2 0.07  0.11
idem in sea water 1.0  1.5 3.3 1.6
aluminium 1.0  2.0 2.9 4.56
aluminium alloy 1.0  2.0 2.6  3.9 3  19
copper 1.8  2.8 3.9 0.34
titanium 2.0  3.0 4.4 68.8
Conversion
Because the dimension of C depends on the value of m, the conversion of units must be done
with care. The example shows the conversion from BS to SI units and vice versa.
da
dN
= C (∆σ
√
πa)
m
→ C =
da
dN
(∆σ
√
πa)
m
[in] and [ksi] → [m] and [MPa]
1
[ in ]
[ ksi
√
in ]
m
=
0.0254 [ m]
{6.86 [ MPa ]
_
0.0254 [ m ] }
m
=
_
0.0254
(1.09)
m
_
[ m]
[ MPa
√
m]
m
[m] and [MPa] → [mm] and [MPa]
1
[ m]
[ MPa
√
m]
m
=
10
3
[ mm]
{[ MPa ]
√
10
3
[
√
mm]}
m
=
_
10
3
{
√
10
3
}
m
_
[ mm]
[ MPa
√
mm]
m
132
12.5.3 Fatigue life
Paris law can be integrated analytically, where a increases from a
i
to a
f
while N goes from
N
i
to N
f
. The result N
f
−N
i
can be represented as a function of a
f
with a
i
as parameter or
vice versa.
N
f
−N
i
=
(∆σ)
−m
β
m
C(
√
π)
m
(1 −
m
2
)
a
(1−
m
2
)
f
_
1 −
_
a
i
a
f
_
(1−
m
2
)
_
Paris law can also be integrated numerically, here, using an explicit integration scheme. The
initial crack length is indicated to be a
0
and the ﬁnal crack length is the critical crack length
a
c
.
set ∆σ, ∆N, a
c
initialize N = 0, a = a
0
while a < a
c
∆K = β ∆σ
√
π ∗ a
da
dN
= C ∗ (∆K)
m
→ ∆a =
da
dN
∗ ∆N
a = a +∆a
N = N +∆N
end
Initial crack length
It is now very easy to show that the initial crack length a
0
is of utmost importance for the
fatigue life. Results are shown for aluminum and for mild steel for a stress amplitude ∆σ = 50
MPa.
0 1 2 3 4 5
x 10
6
0
20
40
60
80
100
120
N [c]
a
[
m
m
]
C = 4.56e−11 ; m = 2.9 ; DN = 100
a0 = 0.1 [mm]
a0 = 1 [mm]
aluminum ; ∆σ = 50 [MPa]
0 2 4 6 8 10 12 14
x 10
6
0
20
40
60
80
100
120
N [c]
a
[
m
m
]
C = 0.25e−11 ; m = 3.3 ; DN = 1000
a0 = 1 [mm]
a0 = 1.1 [mm]
mild steel ; ∆σ = 50 [MPa]
Fig. 12.18 : Fatigue crack length in aluminum and mild steel for two values of the initial
crack length.
133
Fatigue load
The fatigue life is reached, when the crack length becomes critical (a = a
c
=
2γ
π
E
∆σ
2
). For
aluminum, the number of cycles to reach this length has been calculated for various stress
amplitudes.
∆σ [MPa] 25 50 75 100
a
0
[mm] 0.1 0.1 0.1 0.1
a
c
[mm] 56 28 12.5 7
N
f
[c] 35070000 4610000 1366000 572000
0 1 2 3 4
x 10
7
0
10
20
30
40
50
60
N [c]
a
[
m
m
]
Fig. 12.19 : Fatigue crack length in aluminum for various load amplitudes.
12.5.4 Other crack grow laws
Numerous special and extended versions of the basic Paris law have been published.
Erdogan (1963)
da
dN
=
C(1 +β)
m
(∆K −∆K
th
)
n
K
Ic
−(1 +β)∆K
with β =
K
max
+K
min
K
max
−K
min
Broek & Schijve (1963)
da
dN
= CK
2
max
∆K
Forman (1967)
da
dN
=
C(∆K)
n
(1 −R)K
c
−∆K
with R =
K
min
K
max
Donahue (1972)
da
dN
= C(∆K −∆K
th
)
m
with ∆K
th
= (1 −R)
γ
∆K
th
(R = 0)
134
Walker (1970)
da
dN
= C
_
∆K
(1 −R)
n
_
m
with m = 0.4 ; n = 0.5
Priddle (1976)
da
dN
= C
_
∆K −∆K
th
K
Ic
−K
max
_
m
with ∆K
th
= A(1 −R)
γ
and
1
2
≤ γ ≤ 1 [Schijve (1979)]
McEvily & Gr¨ oger (1977)
da
dN
=
A
Eσ
v
(∆K −∆K
th
)
2
_
1 +
∆K
K
Ic
−K
max
_
with ∆K
th
=
_
1 −R
1 +R
∆K
0
A, ∆K
0
∼ inﬂuence environment
NASA / FLAGRO program (1989)
da
dN
=
C(1 −R)
m
∆K
n
(∆K −∆K
th
)
p
[(1 −R)K
Ic
−∆K]
q
m = p = q = 0 → Paris
m = p = 0, q = 1 → Forman
p = q = 0, m = (m
w
−1)n → Walker
12.5.5 Crack growth at low cycle fatigue
Paris law and its extended version are valid for the high cycle fatigue regime, where stresses
are so low that ∆K characterizes the stress amplitude. For low cycle fatigue this is not the
case any more. Crack grow laws for this high stress regime with large plastic crack tip regions,
are still under development.
slipline
θ
σ
λσ
θ
Fig. 12.20 : Crack under multimode fatigue loading.
135
da
dN
=
3 −sin
−2
(θ) cos
−2
(
θ
2
)
9 sin(θ)
K
Eσ
v
_
1 −βγ
−
1
2
_
K
2
max
{1 −(1 −λ)
σmax
σv
}
θ = cos
−1
_
1
3
_
β
√
γ
= 0.5 + 0.1R + 0.4R
2
_
_
_
→
da
dN
=
7
64
√
2
K
Eσ
v
(1 −0.2R −0.8R
2
)
K
2
max
{1 −(1 −λ)
σmax
σv
}
For high values of crack tip stress and consequently a large plastic zone, the Paris law may
be used with ∆J instead of ∆K. Note that ∆J = J
max
−J
min
over a cycle.
da
dN
= C
∗
(∆J)
m
∗
with ∆J =
_
Γ
_
W
∗
n
1
−∆t
i
∂∆u
i
∂x
1
_
dΓ ; W
∗
=
εpqmax
_
εpq
min
∆σ
ij
dε
ij
12.6 Load spectrum
Until now, we only looked at harmonic loads with constant frequency and amplitude. In prac
tical situations, such loads will almost never be seen. Both loading frequency and amplitude
will generally vary over time. Cyclic loading with such a time dependent character is referred
to as a load spectrum. It is assumed here that the direction of the load remains the same.
A load spectrum is shown in the ﬁgure below. It consists of four (i = 1..4) harmonic
loads with diﬀerent frequencies and amplitudes. In the subsequent loads the number of cycles
is n
i
. For each individual harmonic load the life time N
if
is known.
Miner has found in 1945 that the fatigue life for the whole spectrum is independent on
the actual sequence of the individual loadings [47]. This means that the individual harmonic
loadings do not inﬂuence the damage growth in the following loading stage with a diﬀerent
character. The ﬁnal life time is reached, when the summation of
n
i
N
if
becomes one. This
socalled Miner’s rule is approximately valid for a wide range of materials. In reality the
ultimate sum of all the ratios is between 0.6 and 2.0, which indicates that there is actual
some interaction between the diﬀerent load cycles.
136
n
4
n
3
n
2
n
1
N
0
σ
Fig. 12.21 : Sequence of various harmonic loadings.
L
i=1
n
i
N
if
= 1
Miner’s rule is clearly illustrated by regarding the rest life after each load cycle. When it
reaches zero, the summation of the ratio
n
i
N
if
must reach the value one.
1 → 1 −
n
1
N
1f
2 →
_
1 −
n
1
N
1f
_
−
n
2
N
2f
3 →
_
1 −
n
1
N
1f
−
n
2
N
2f
_
−
n
3
N
3f
4 →
_
1 −
n
1
N
1f
−
n
2
N
2f
−
n
3
N
3f
_
−
n
4
N
4f
= 0
12.6.1 Random load
Although the load spectrum is already much more realistic than the harmonic loading with
constant frequency and amplitude, practical loading is mostly rather random, as is illustrated
in the ﬁgure below.
Prediction of fatigue life is only possible after this random load is transferred into a load
spectrum, a sequence of harmonic loads, with known frequencies and amplitudes. Various
counting procedures can be used for this transformation. Literature gives detailed information
about (mean crossing) peak count, range pair (mean) count and rain ﬂow count techniques.
137
0
t
σ
Fig. 12.22 : Random load.
For some applications, random load sequences are measured and used in design. During
service life the load can be monitored and the expected fatigue life time can be updated.
Typical loadings for a manufacturing machine (a), a sea vessel (b) and an airplane (c) are
shown in the ﬁgure.
138
Fig. 12.23 : Random loads for machine (a), sea vessel (b) and airplane (c).
12.6.2 Tensile overload
Known random loadings are used to calculate the fatigue life during the design of structures.
During service life extreme overloads can occur, which must of course be low enough to cause
no direct damage. The question is how they aﬀect the fatigue life.
Although the intuitive feeling is that overload must be bad, practical experience shows
that the crack propagation is slowed down by such an overload, at least during a limited
number of cycles. This is illustrated in the next ﬁgure, where (K
max
)
OL
is the maximum SIF
during the overload.
139
N
a
K
max
(K
max
)
OL
b
1
b
2
a
Fig. 12.24 : Overload and crack retardation.
The crack retardation is also illustrated quantitatively by the values in the next table (from
[29]), which apply to the Mode I loading of specimen made of Al 2024T3 alloy.
∆K % P
max
nr. P
max
delay
[MPa
√
m] [] [] [10
3
cycles]
15 53 1 6
15 82 1 16
15 109 1 59
16.5 50 1 4
16.5 50 10 5
16.5 50 100 9.9
16.5 50 450 10.5
16.5 50 2000 22
16.5 50 9000 44
23.1 50 1 9
23.1 75 1 55
23.1 100 1 245
The crack retardation can be explained by regarding the stress state in a region in front of
the crack tip. The ﬁgure shows the stress σ
yy
and the plastic limit stress. In a certain time
period, the loading stress is increased to a maximum value σ
1
, representing the overload, and
140
decreased again. In two points A and B the stress σ
yy
is plotted in the ﬁgure as a function of
the strain ε
yy
. In point A, the material is plastically deformed, while the deformation in point
B remains elastic. After unloading, the strain in A and B is more or less the same, so that it
becomes clear from the stressstrain curves that a compressive residual stress remains in point
A. This compressive stress inhibits the opening of the crack and thus the crack elongation.
When the crack tip has moved through the compressive zone, the normal crack grow rate is
again established.
B A
σ
v
σ
0
t
σ
σ
1
σ
σ
yy
σ = 0
σ
yy
B
1
A
1
A
2
σ
yy
ε
yy
B
2
Fig. 12.25 : Residual stress at the crack tip resulting from overload.
Crack retardation models
Willenborg (1971) has published a relation for the SIF, K
R
, in the crack retardation stage.
It has been enhanced by Johnsson [37]. Both relations are taking into account the inﬂuence
of the compressive stress zone at the front of the crack tip.
141
Willenborg (1971)
K
R
= φ
_
(K
max
)
OL
_ ¸
1 −
∆a
r
y
_
−K
max
_
; ∆a < r
y
K
R
= residual SIF ; K
R
= 0 → delay distance
φ = [1 −(K
th
/K
max
)](S −1)
−1
; S = shutoﬀ ratio
r
y
a
K
max
(K
max
)
OL
∆a
Johnson (1981)
R
eff
=
K
min
−K
R
K
max
−K
R
; r
y
=
1
βπ
_
(K
max
)
OL
σ
v
_
2
β = plastic constraint factor
Crack retardation due to crack closure, caused by the compressive stress zone behind the
crack tip, have also been devised and published.
Elber (1971) [17] ∆K
eff
= U ∆K ; U = 0.5 + 0.4R with −0.1 ≤ R ≤ 0.7
Schijve (1981) [58] U = 0.55 + 0.33R + 0.12R
2
with −1.0 < R < 0.54
12.7 Design against fatigue
There are three diﬀerent strategies which can be followed to prevent that fatigue damage
leads to failure, socalled design against fatigue.
Inﬁnite life design
For inﬁnite life design, the stress amplitude in the most severely loaded parts of the structure
are kept below the fatigue limit. The result is that fatigue crack growth will be practically
zero. This strategy will involve the use of much material and thus lead to bulky and heavy
structures.
142
Safe life design
Using the concept of safe life, the (SN)curve is used to predict when the fatigue life is
reached. For this purpose, the loading must be known and the number of cycles must be
logged.
When the load is random, satistical counting procedures must be employed to translate
the random load spectrum to a sequence of harmonic loads.
Instead of the (SN)curve, the ﬁtted Basquin or MansonCoﬃn relations can be used. They
can be combined in a stress/strain life curve.
When the number of cycles to failure – probable a much lower safe number – is reached,
the structur(e)(al) (part) is replaced.
Damage tolerant design
Crack growth can be predicted, using a crack growth relation like Paris law. Periodic inspec
tion is of course needed to check whether predicted crack lengths are correct. The rest life
can be calculated, including eﬀects like overloads. When the critical crack length – probable
a much smaller value – is reached, the structur(e)(al) (part) is replaced.
Fail safe design
Although safety factors will certainly be used to be within safe margins of the fatigue life,
The design must be such that (undetected) fatigue cracks will not lead to global failure.
Chapter 13
Engineering plastics (polymers)
Polymer materials are widely used in functional and structural applications. External loads
and other inﬂuences lead of course to damage, cracking and failure. Systematic study of these
phenomena has brought insight in their behavior and, even more than in the case of metal
alloys, has given rise to the development of new materials with better damage resistance.
There are a huge number of diﬀerent polymer materials, some of which are listed in the
table.
ABS acrylonitrilbutadieenstyreen EM
HIPS highimpact polystyreen TP
LDPE lowdensity polyetheen TP
Nylon
PC polycarbonaat TP
PMMA polymethylmethacrylaat (plexiglas) TP
PP polypropeen TP
PPO polyfenyleenoxyde TP
PS polystyreen TP
PSF PolySulFon TP
PTFE polytetraﬂuoretheen (teﬂon) TP
PVC polyvinylchloride TP
PVF polyvinylﬂuoride TP
PVF2 polyvinylidieenﬂuoride TP
13.1 Mechanical properties
For many applications, metal alloys show linear elastic behavior, in which case Linear Elastic
Fracture Mechanics can be used to study cracks. Polymers mostly have a lower stiﬀness and
the linear elasticity limit is reached sooner. Polymer materials generally show strainrate
and timedependent behavior. Temperature eﬀects cannot be neglected in many cases. In
fracture analysis of polymers it is therefore often needed to use concepts from NonLinear
Fracture Mechanics. Composite materials are anisotropic, which, even in the linear elastic
case, complicates the mechanical behavior.
143
144
13.1.1 Damage
Damage in polymers may occur as shearbands or crazes. These can occur both in the same
material in regions with diﬀerent stress states.
Shear bands are associated to shear yielding. In this case no change in density is observed.
Crazes are connected to normal yielding. The local density may decrease with 40  60 %. In
a craze, ﬁbrils are connecting the two craze surfaces. When the craze opens, the ﬁbrils are
elongated due to extrusion of macromolecules from the bulk polymer. When the maximum
elongation is reached, the ﬁbril snaps.
crazes shear bands
Fig. 13.1 : Shear bands and crazes in a polymer.
13.1.2 Properties of engineering plastics
The next table shows some properties of engeneering plastics, which are related to their
damage behavior. The crystallinity is a very important item, because crystals can stop cracks
in an early stage. The occurence of crazes is also relevant, because crazing is generally
associated to brittle behavior.
145
AC CZ K
Ic
PMMA a + 13.2
PS a + 17.6
PSF a  low
PC a  high main chain segmental motions →
energy dissipation
Nylon 66 cr  main chain segmental motions →
energy dissipation
crystalline regions → crack retardation
PVF2 sc
PET sc amorphous →
strain induced crystallization at crack tip
CPLS  crosslinked → suppressed crazing
HIPS + µsized rubber spheres →
enhanced crazing
ABS blending
AC : a = amorphous
AC : c = crystalline
AC : sc = semicrystalline
AC : cr = crystalline regions
CZ = crazing
K
Ic
= fracture toughness in MPa
√
m
13.2 Fatigue
Fatigue crack propagation in polymers is strongly aﬀected by properties like molecular weight,
cross linking, main chain mobility and crystallinity. Blending one polymer with another
polymer or particles to improve ductility and fatigue life is referred to as toughening. Data
for the next ﬁgures are taken from [29].
146
0.5
10
−4
[mm/cycle]
3 ∗ 10
−4
10
−3
3 ∗ 10
−3
10
−2
5 2 1
1
2
3
4
5
da
dN
[MPa
√
m] ∆K
1 : PMMA 5 Hz crazing
2 : LDPE 1 Hz
3 : ABS 10 Hz
4 : PC 10 Hz no crazing
5 : Nylon 10 Hz cristalline regions
0.5
9
1.0 2.0 5.0 10.0
10
−5
10
−4
10
−3
10
−2
10
−1
1
2
3
4
5
6
7
[mm/cycle]
da
dN
∆K [MPa
√
m]
8
1 : PS 5 : PC
2 : PMMA 6 : Nylon 6.6
3 : PSF 7 : PVF2
4 : PPO 8 : PET
9 : PVC
147
4
da
dN
3
2
1
10
−2
10
−3
10
−4
5.0 2.0 1.0 0.5
∆K [MPa
√
m]
[mm/cycle]
1 : CLPS
2 : PS
3 : HIPS
4 : ABS
148
Chapter 14
Cohesive zones
With ongoing miniaturization, interfaces get more important concerning the (mechanical)
behavior of material systems. The degradation of interfaces – e.g. delamination – can be
studied and analyzed with cohesive zone models.
These models are used to investigate the behavior of interfaces between polymer coating
and steel sheet substrate, a material system used more and more for all kind of applications.
Also the peeloﬀ lid of food containers is relying on the proper delamination of a polymer
interface layer. As a ﬁnal example, the interfaces in solder joints for microelectronic compo
nents are studied, with special attention for their fatigue life.
The next sections are based on the research work of Marco van den Bosch (PhD), Sil
Bijker (MSc) and M¨ uge Erin¸ c (PhD), who’s papers and reports are listed in the reference
section [].
14.1 Cohesive zone models
Cohesive zone (cz) models have been introduced by Dugdale and Barenblatt [15] and have
attracted a growing interest in the scientiﬁc community to describe failure processes and
delamination in particular. Cohesive zones project all damage mechanisms in and around a
crack tip on the interface, leading to a constitutive relation, or cohesive zone law. Interaction
between crack faces is automatically incorporated and can be ﬁtted on experimental data.
Cohesive zone models relate the relative displacement (”opening” ∆) of two associated
points of the interface to the force per unit of area (”traction” T) needed for separation.
Frequently – but not necessarily – a diﬀerence is made between normal (n) and tangential (t)
direction, so the cohesive zone law comprises the two relations T
n
(∆
n
) and T
t
(∆
t
).
Cohesive zone laws can be uncoupled or coupled. In an uncoupled cohesive zone law the
normal/tangential traction is independent of the tangential/normal opening. In a coupled
cohesive zone law, both normal and tangential tractions depend on both the normal and
tangential opening displacement. Uncoupled laws are intended to be used when the debonding
process occurs under one mode – normal (modeI, opening) or tangential (modeII, sliding)
loading – or is largely dominated by one mode. The majority of cohesive zone laws have a
(partial) coupling between normal and tangential direction, which is achieved by introducing
coupling parameters in the model.
149
150
t
n
Fig. 14.1 : An undeformed and deformed interface with a schematic illustration of a
cohesive zone tractionopening interaction in normal and tangential direction.
traction = f(opening) T = f(∆)
normal / tangential T
n
= f
n
(∆
n
) ; T
t
= f
t
(∆
t
)
coupling T
n
= f
n
(∆
n
, ∆
t
) ; T
t
= f
t
(∆
t
, ∆
n
)
A large variety of cohesive zone laws has been described in literature. Most of them can be
categorized into the following groups: polynomial, piecewise linear, rigidlinear and expo
nential. These cohesive zone laws are described in the next sections. The maximum normal
traction and the maximum tangential traction are indicated by T
n,max
and T
t,max
, respec
tively. Characteristic openings are δ
n
and δ
t
. The areas below the tractionopening curves
represent the normal and tangential workofseparation W
n
and W
t
.
14.1.1 Polynomial
In the polynomial cohesive zone law, coupling is achieved with an eﬀective opening displace
ment λ and a coupling parameter α, which relates the maximum tractions. The function f(λ)
is such that the tractions are nonzero for 0 < λ ≤ 1 and remain zero for λ > 1. The ﬁgure
below illustrates this cohesive zone behavior with the function f(λ) taken from [64]. The
coupling is incorporated well. The plots show that the tractions in one direction deminish,
when the cohesive zone is opened in the other direction. The largest tractions are associated
with the opening in one direction or, which is the same, with an uncoupled formulation. Note
that in this cohesive zone δ
n
and δ
t
indicate the opening, when the traction is zero.
T
n
= T
n,max
∆
n
δ
n
f(λ; 0 < λ ≤ 1)
T
t
= T
t,max
∆
t
δ
t
f(λ; 0 < λ ≤ 1) = αT
n,max
∆
t
δ
t
f(λ; 0 < λ ≤ 1)
eﬀective displacement λ =
¸
_
∆
n
δ
n
_
2
+
_
∆
t
δ
t
_
2
function from [64] f(λ) =
27
4
_
1 −2λ +λ
2
_
151
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.2 : Polynomial cohesive zone law with coupling.
14.1.2 Piecewise linear
In the piecewise linear cohesive zone law, shown below, the tractionopening function is
trapezoidal. No coupling is taken into account here. Due to the assumed relation between
the maximum tractions, the normal and tangential workofseparation have the same value.
T
t,max
= (δ
n
/δ
t
)T
n,max
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.3 : Piecewise linear cohesive zone law without coupling.
14.1.3 Rigidlinear
Cohesive zones may be modelled to have a very high stiﬀness, until the maximum traction
is reached. In such a rigidlinear cohesive zone law, as proposed in [12], the normal traction
is not a function of the tangential opening ∆
t
, while the tangential traction is controlled
completely by the normal opening ∆
n
. This implies that the interface can only fail due
to decohesion in normal direction. The plot shows a rigidlinear law without this coupling
(α = 1). The characteristic openings δ
n
and δ
t
are the openings at zero traction.
152
T
n
= T
n,max
_
1 −
∆
n
δ
n
_
∆
n
> 0
T
t
= T
t,max
_
1 −α
∆
t
δ
t
_
∆
t
> 0
_
¸
¸
_
¸
¸
_
α =
δ
t
∆
t
∆
n
δ
n
coupling
α = 1 no coupling
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.4 : Rigidlinear cohesive zone law without coupling.
14.1.4 Exponential
The exponential cohesive zone law is used very much. It has some advantages compared to
other laws. First of all, a phenomenological description of contact is automatically achieved
in normal compression. Secondly, the tractions and their derivatives are continuous, which is
attractive from a computational point of view. The exponential cohesive zone law originates
from the universal relationship between binding energies and atomic separation of interfaces.
Two exponential czlaws are described in the following, both based on a potential φ,
which is a function of both the normal and the tangential opening. The ﬁrst law, proposed by
Xu & Needleman is much used, but appears to have some deﬁciancies in describing coupling.
Xu & Needlemaan exponential czlaw
The Xu&Needleman cohesive zone law [69] is deﬁned by one potential φ which depends on
the normal and tangential openings ∆
n
and ∆
t
. There are two coupling parameters, q and r,
which will be discussed below.
φ(∆
n
, ∆
t
) = φ
n
+φ
n
exp
_
−
∆
n
δ
n
_
__
1 −r +
∆
n
δ
n
__
1 −q
r −1
_
−
_
q +
_
r −q
r −1
_
∆
n
δ
n
_
exp
_
−
∆
2
t
δ
t
2
__
Tractions in normal (T
n
) and tangential (T
t
) direction are derived from the potential φ by
diﬀerentiation w.r.t. ∆
n
and ∆
t
respectively.
153
T
n
=
∂φ
∂∆
n
=
φ
n
δ
n
exp
_
−
∆
n
δ
n
__
∆
n
δ
n
exp
_
−
∆
2
t
δ
t
2
_
+
1 −q
r −1
_
1 −exp
_
−
∆
2
t
δ
t
2
___
r −
∆
n
δ
n
__
T
t
=
∂φ
∂∆
t
= 2
φ
n
δ
t
∆
t
δ
t
_
q +
_
r −q
r −1
_
∆
n
δ
n
_
exp
_
−
∆
n
δ
n
_
exp
_
−
∆
2
t
δ
t
2
_
For pure normal and tangential loading, the tractionopening relations are simpliﬁed. In
troducing the parameter φ
t
, the coupling parameter q is deﬁned as q = φ
t
/φ
n
. The nor
mal and tangential tractionopening relations are plotted for the parameter values φ
n
=
100 Jm
−2
, φ
t
= 80 Jm
−2
, T
n,max
= 10 N , T
t,max
= 10 N .
T
n
= T
n
(∆
t
= 0) =
φ
n
δ
n
∆
n
δ
n
exp
_
−
∆
n
δ
n
_
T
t
= T
t
(∆
n
= 0) = 2
φ
n
δ
t
∆
t
δ
t
q exp
_
−
∆
2
t
δ
t
2
_
= 2
φ
t
δ
t
∆
t
δ
t
exp
_
−
∆
2
t
δ
t
2
_
T
n,max
= T
n
(∆
n
= δ
n
) =
φ
n
δ
n
exp(1)
→ δ
n
=
φ
n
T
n,max
exp(1)
T
t,max
= T
t
(∆
t
=
_
1
2
δ
t
) =
φ
t
δ
t
_
1
2
exp(1)
→ δ
t
=
φ
t
T
t,max
_
1
2
exp(1)
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.5 : Uncoupled normal and tangential tractionopening for the Xu&Needleman
cohesive zone law.
When the cohesive zone is loaded simultaneously (or subsequently) in normal and tangential
direction, the coupling is taken into account. The results are dependent of the value of r. For
r = 0 the normal traction does not become zero, when the interface is failed in tangential
direction. For r = q and other values of r, the results are even more unrealistic, as the normal
traction is not zero for zero normal opening.
154
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.6 : Coupled normal and tangential tractionopening for the Xu&Needleman
cohesive zone law for r = 0.
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.7 : Coupled normal and tangential tractionopening for the Xu&Needleman
cohesive zone law for r = q.
Adapted exponential cohesive zone law
An exponential cohesive zone law, which does not have the anomalities of the Xu&Needleman
law, is again based on a potential [5]. The potential incorporates four independent parameters:
the workofseparation for pure normal opening, φ
n
, the workofseparation for pure tangential
opening, φ
t
, the characteristic opening in normal direction, δ
n
, and the characteristic opening
in tangential direction, δ
t
.
The normal and tangential tractions are calculated by diﬀerentiating the potential w.r.t.
the normal and tangential opening, respectively. For pure normal and tangential opening, the
tractions are plotted in the ﬁgure below and they are identical to those of the Xu&Needleman
law. The characteristic lengths δ
n
and δ
t
can be expressed in the maximum tractions.
φ(∆
n
, ∆
t
) = φ
nt
_
1 −
_
1 +
∆
n
δ
n
_
exp
_
−
∆
n
δ
n
_
exp
_
−
∆
2
t
δ
t
2
__
T
n
=
∂φ
∂∆
n
¸
¸
¸
¸
φn
=
φ
n
δ
n
_
∆
n
δ
n
_
exp
_
−
∆
n
δ
n
_
exp
_
−
∆
2
t
δ
t
2
_
T
t
=
∂φ
∂∆
t
¸
¸
¸
¸
φt
= 2
φ
t
δ
t
_
∆
t
δ
t
__
1 +
∆
n
δ
n
_
exp
_
−
∆
2
t
δ
t
2
_
exp
_
−
∆
n
δ
n
_
155
δ
n
=
φ
n
T
n,max
exp(1)
; δ
t
=
φ
t
T
t,max
_
1
2
exp(1)
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.8 : left: The uncoupled normal response of the exponential cohesive zone law with
the maximum traction T
n,max
at ∆
n
/δ
n
= 1. right: The uncoupled tangential response of
the exponential cohesive zone law with the maximum traction T
t,max
at ∆
t
/δ
t
= 1/
√
2.
Coupling
Adequate coupling between the normal and tangential directions is required in a cohesive
zone law to describe the physically occurring interface behavior realistically. If complete
loss of interfacial integrity is important (e.g. in the case of a moving delamination front)
and a cohesive zone completely fails in shear, its loadcarrying capacity in normal traction
should completely vanish as well and vise versa. That this behavior is described adequately,
is illustrated by the tractionopening plots below.
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
T
n
/
T
n
,
m
a
x
∆n/δn
−2 −1 0 1 2
−1
−0.5
0
0.5
1
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.9 : The coupled normal and tangential response of the exponential cohesive zone
law.
156
To investigate this behavior in more detail, the interface is ﬁrst loaded in normal direction
until a maximum opening ∆
n,max
. After that it is broken in shear : ∆
t
→∞, as is illustrated
in the ﬁgure below. The shaded areas under the curves are the work of separation for normal
and tangential opening.
W
n
=
∆n,max
_
0
T
n
(∆
n
)
∆t=0
d∆
n
; W
t
=
∞
_
0
T
t
(∆
t
)
∆n=∆n,max
d∆
t
Fig. 14.10 : Loading sequence to study the inﬂuence of coupling parameters on
workofseparation: (left) interface loaded to ∆
n,max
; (right) interface broken in shear
∆
t
→ ∞.
Several loading sequences are evaluated. In the left ﬁgure below the evolution of the maximum
shear traction is shown as a function of the normal separation. As can be seen in this ﬁgure, the
maximum shear traction decreases to zero for increasing normal separations. Theoretically,
this can also happen in compression. However, these negative values of ∆
n
cannot be reached
in practice since the normal compressive traction increases very fast in this regime.
In the right ﬁgure below the maximum normal traction as a function of the tangential
separation is shown. The maximum normal traction decreases to zero for increasing tangential
separation.
T
∗
t
= max{T
t
(∆
t
, ∆
n,max
)}
−1 0 1 2 3 4
0
0.2
0.4
0.6
0.8
1
T
∗ t
/
T
t
,
m
a
x
∆n,max/δn
T
∗
n
= max{T
n
(∆
n
, ∆
t,max
)}
−2 −1 0 1 2
0
0.2
0.4
0.6
0.8
1
T
∗ n
/
T
n
,
m
a
x
∆t,max/δt
Fig. 14.11 : left: The normalized maximum shear traction as a function of the normalized
normal separation of the alternative exponential law.
right: The normalized maximum normal traction as a function of the normalized tangential
separation of the alternative exponential law.
157
Coupling should also be realistic, when considering the total workofseparation W
tot
= W
n
+
W
t
in case of a sequential loading, in normal and tangential direction or viceversa..
The behavior is as expected and shown in the ﬁgures below : in the case where there is
ﬁrst normal separation until ∆
n,max
and then complete tangential separation, W
tot
increases
monotonically from the value φ
t
to the value of φ
n
.
In the other case where there is ﬁrst tangential separation until ∆
t,max
and then complete
normal separation, W
tot
smoothly decreases from the value of φ
n
to φ
t
.
0 2 4 6 8
0
20
40
60
80
100
W
[
J
m
−
2
]
W
n
W
t
W
tot
∆n,max/δn
0 2 4 6 8
0
20
40
60
80
100
W
[
J
m
−
2
]
W
n
W
t
W
tot
∆t,max/δt
Fig. 14.12 : left: Workofseparation when the cohesive zone is ﬁrst loaded in normal
direction until ∆
n,max
and then broken in shear. right: Workofseparation when the
cohesive zone is ﬁrst loaded in tangential direction until ∆
t,max
and then broken in tension.
Modemixity
The inﬂuence of mixed mode loading on the total workofseparation is studied next. To this
purpose, a cohesive zone is loaded under an angle α w.r.t. the normal direction until complete
separation occurs, and the corresponding value of W
tot
is calculated.
W
tot
is shown as a function of the loading angle α. Again, values φ
n
= 100 Jm
−2
and
φ
t
= 80 Jm
−2
are used in these calculations. In the ﬁgure below it can be seen that W
tot
= φ
n
when the cohesive zone is only loaded in normal direction (modeI: α = 0
o
) and W
tot
= φ
t
when it is only loaded in tangential direction (modeII: α = 90
o
).
14.2 Weighted residual formulation with cohesive zones
In the deformed state of a material body, the equations of motion must be satisﬁed in each
material point with position x in the current material volume V . When material acceleration
is suﬃciently small, the righthand side is zero and the equilibrium equation results, a partial
diﬀerential equation, which can be solved with proper boundary conditions.
∇· σ +ρ q = ρ
˙
v ∀ x ∈ V
eq. of equilibrium
∇· σ +ρ q =
0 ∀ x ∈ V
158
with
∇ : gradient operator
σ : Cauchy stress tensor
ρ : density
q : volume load
v : material velocity
Analytical solutions are seldomly reached for practical problems and we have to be satisﬁed
with approximate solutions, which can be determined with the ﬁnite element method. To this
purpose the diﬀerential equation is transformed into a weighted residual integral by use of a
vectorial weighting function w(x). Partial integration of the ﬁrst term leads to a weak version.
The two integrals with the external loads are abbreviated as the external load integral f
e
.
The remaining integral is the internal load integral f
i
.
∇· σ +ρ q =
0 ∀ x ∈ V ⇐⇒
_
V
w·
_
∇· σ +ρ q
_
dV = 0 ∀ w(x)
_
V
(
∇ w)
c
: σ dV =
_
V
w · ρ q dV +
_
A
w ·
t dA ∀ w(x)
f
i
( w, σ) = f
e
( w,
t, q) ∀ w(x)
The internal load integral can be written as the sum of an integral over bulk material (volume
V
b
) and an integral over cohesive zones (volume V
cz
).
f
i
=
_
V
(
∇ w)
c
: σdV =
_
V
b
(
∇ w)
c
: σ dV +
_
Vcz
(
∇ w)
c
: σ dV = f
i
b
+f
icz
When the weighting function w is chosen to be a virtual displacement δu, the weighted
residual integral becomes the virtual work equation in which the logarithmic strain ε
ln
is
introduced. We only consider the integral over the cohesive zones. Furthermore, we consider
a twodimensional case, where the cz has uniform depth d. The unity vector e is deﬁned along
the line between two associated points of the interface. The deformed opening of the cohesive
zone is h.
f
icz
=
_
Vcz
(
∇ w)
c
: σ dV
w = δu → (
∇ w)
c
: σ = (
∇δu)
c
: σ = δε
ln
: σ
f
icz
=
_
Vcz
δε
ln
: σ dV = d
_
Acz
δε
ln
: σ dA
δε
ln
= δε
ln
ee = δε
ln
e
σ = σee = σe
_
→ δε
ln
: σ = δε
ln
· σ
f
icz
= d
_
Acz
δε
ln
· σ dA = d
_
Acz
δ∆u
h
· σ dA = d
_
Acz
∆(δu)
h
· σ dA
=
d
h
_
Acz
∆ w· σ dA
159
When large deformations occur, the integral has to be transformed to a known state. In this
case we choose the initial undeformed state. In this transformation the deformation tensor F
is used. The transformed integral contains the ﬁrst PiolaKirchhoﬀ stress tensor T.
_
V
(
∇ w)
c
: σdV = f
e
( w,
t, q) ∀ w(x)
∇ = F
−c
·
∇
0
→ (
∇ w)
c
= (
∇
0
w)
c
· F
−1
dV = det(F) dV
0
= J dV
0
_
V
0
(
∇
0
w)
c
· F
−1
: σJ dV
0
= f
e0
∀ w(x)
_
V
0
(
∇
0
w)
c
: (F
−1
· σJ) dV
0
= f
e0
∀ w(x)
_
V
0
(
∇
0
w)
c
: T dV
0
= f
e0
∀ w(x)
f
i0
( w, T) = f
e0
( w,
t
0
, q
0
) ∀ w(x)
Again the internal load integral is written as sum of an integral over the bulk volume and an
integral over the cohesive zones.
f
i0
=
_
V
0
(
∇
0
w)
c
: T dV
0
=
_
V
0
b
(
∇
0
w)
c
: T dV
0
+
_
V
0cz
(
∇
0
w)
c
: T dV
0
= f
i0
b
+f
i0cz
When the weighting function w is chosen to be a virtual displacement δu, the weighted
residual integral becomes the virtual work equation in which the linear strain is introduced.
We only consider the integral over the cohesive zones and take the cz to have a uniform initial
depth d
0
. The unity vector e
0
is deﬁned along the line between two associated points of the
interface. The undeformed opening of the cohesive zone is h
0
.
f
i0cz
=
_
V
0cz
(
∇
0
w)
c
: T dV
0
w = δu → (
∇
0
w)
c
: T = (
∇
0
δu)
c
: T = δε
l
: T
f
i0cz
=
_
V
0cz
δε
l
: T dV
0
= d
0
_
A
0cz
δε
l
: T dA
0
δε
l
= δε
l
e
0
e
0
= δε
l
e
0
T = T e
0
e
0
=
T e
0
_
→ δε
l
: T = δε
l
·
T
f
i0cz
= d
0
_
A
0cz
δε
l
·
T dA
0
= d
0
_
A
0cz
δ∆u
h
0
·
T dA
0
= d
0
_
A
0cz
∆(δu)
h
0
·
T dA
0
=
d
0
h
0
_
A
0cz
∆ w·
T dA
0
160
14.3 Twodimensional CZ element
A cohesive zone element has uniform initial thickness d
0
. The length l
0
of the element is
deﬁned to be the undeformed length of the line AB between the midpoints of the element
edges 14 and 23. Along this line a local coordinate η is introduced, which spans the range
[−1, 1].
In the deformed state a traction
T works between associated (= with the same η
coordinate) points P and Q on edges 12 and 43 respectively. This traction is a function of
the elongation of the line PQ w.r.t. the undeformed state, given by the cohesive zone law.
Along PQ the local coordinate ξ is deﬁned in a range [−1, 1].
1
1
B
4
3
1
2
A
Q
P
ξ
T
η
0
f
i0cz
=
d
0
h
0
_
A
0cz
∆ w·
T dA
0
=
d
0
h
0
_
1
ξ=−1
_
1
η=−1
∆ w·
T
h
0
2
l
0
2
dηdξ
=
d
0
2
_
1
ξ=−1
_
1
η=−1
∆ w·
T
l
0
2
dηdξ =
d
0
l
0
2
_
1
η=−1
∆ w·
T dη
14.3.1 Local vector base
The weighting function and the traction are written in components w.r.t. a local orthonormal
basis {e
t
, e
n
}, which is deﬁned in the midpoint of AB.
4
3
1
T
2
A
B
e
t
T
e
n
∆ w = ∆w
t
e
t
+ ∆w
n
e
n
= [∆w
t
∆w
n
]
_
e
t
e
n
_
= ∆w
˜
T
e
˜
T = T
t
e
t
+T
n
e
n
=
_
T
t
T
n
¸
_
e
t
e
n
_
= T
˜
T
e
˜
f
i0cz
=
d
0
l
0
2
_
1
η=−1
∆w
˜
T
(η)T
˜
(η) dη
161
Interpolation
The weighting function components are interpolated between their values in point A and
point B. The interpolation functions are linear in the local coordinate.
∆w
˜
T
(η) = [∆w
t
∆w
n
] ==
_
∆w
A
t
∆w
A
n
∆w
B
t
∆w
B
n
¸
_
¸
¸
_
1
2
(1 −η) 0
0
1
2
(1 −η)
1
2
(1 +η) 0
0
1
2
(1 +η)
_
¸
¸
_
= ∆w
˜
AB
T
N
T
(η)
The values ∆w in points A and B are written as the diﬀerences of the weighting function in
the nodal points 4 and 1 and 3 and 2 respectively. The index l indicates that components are
taken w.r.t. the local vector basis {e
t
, e
n
}.
∆w
A
t
= w
4
t
−w
1
t
; ∆w
A
n
= w
4
n
−w
1
n
∆w
B
t
= w
3
t
−w
2
t
; ∆w
B
n
= w
3
n
−w
2
n
A matrix P is introduced to express ∆w
˜
AB
in the local nodal displacements in the column
w
˜
l
.
∆w
˜
AB
=
_
¸
¸
¸
_
∆w
A
t
∆w
A
n
∆w
B
t
∆w
B
n
_
¸
¸
¸
_
=
_
¸
¸
¸
_
−1 0 0 0 0 0 1 0
0 −1 0 0 0 0 0 1
0 0 −1 0 1 0 0 0
0 0 0 −1 0 1 0 0
_
¸
¸
¸
_
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
w
1
t
w
1
n
w
2
t
w
2
n
w
3
t
w
3
n
w
4
t
w
4
n
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
= P w
˜
l
The column with weighting functions is transformed from the local to the global coordinate
system, which is done with a rotation matrix R. It transforms the local base vectors {e
t
, e
n
}
to the global base vectors {e
x
, e
y
}. The only nonzero components in the rotation matrix
are cosine (c) and sine (s) functions of the angle between the line AB and the global xaxis.
The cosine and sine values can be easily calculated from the coordinates of points A and B :
c = (x
B
−x
A
)/ℓ and s = (y
B
−y
A
)/ℓ, where ℓ is the deformed length of line AB. Coordinates
of A and B can be expressed in the nodal point coordinates : xy
AB
=
1
2
(xy
12
+xy
43
).
w
˜
l
=
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
w
1
t
w
1
n
w
2
t
w
2
n
w
3
t
w
3
n
w
4
t
w
4
n
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
=
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
c s 0 0 0 0 0 0
−s c 0 0 0 0 0 0
0 0 c s 0 0 0 0
0 0 −s c 0 0 0 0
0 0 0 0 c s 0 0
0 0 0 0 −s c 0 0
0 0 0 0 0 0 c s
0 0 0 0 0 0 −s c
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
w
1
1
w
1
2
w
2
1
w
2
2
w
3
1
w
3
2
w
4
1
w
4
2
_
¸
¸
¸
¸
¸
¸
¸
¸
¸
¸
_
= Rw
˜
162
So we can write :
∆w
˜
T
= w
˜
T
l
P
T
N
T
(η) = w
˜
T
R
T
P
T
N
T
(η)
and
f
i0cz
= w
˜
T
d
0
ℓ
0
2
R
T
P
T
_
1
η=−1
N
T
(η)T
˜
(η) dη = w
˜
T
f
˜
i0cz
(R, T
˜
)
After assembling all elements, the global weighted residual equation has to be satisﬁed for all
nodal values of the weighting function. This means that the nodal point internal forces have
to be in equilibrium with the external nodal forces.
w
˜
T
f
˜
i0
b
+w
˜
T
f
˜
i0cz
= w
˜
T
f
˜
e0
∀ w
˜
f
˜
i0
b
(σ, F) +f
˜
i0cz
(R, T
˜
) = f
˜
e0
14.3.2 Iterative procedure
The internal load column f
˜
i0cz
is a nonlinear function of the traction T
˜
and the element
rotation R. The unknown opening of the cohesive zone must be solved from nonlinear (global)
equations in an iterative procedure, which means that each unknown quantity is written as
its approximated value ( )
∗
and an iterative change δ( ). This is substituted in the internal
load column, which is linearized subsequently. It is assumed here that δR ≈ 0. The iterative
changes of the traction can be expressed in the iterative opening.
T
˜
= T
˜
∗
+δT
˜
; R = R
∗
+δR ≈ R
∗
→ f
˜
i
= f
˜
∗
i
+δf
˜
i
δf
˜
i0cz
=
d
0
ℓ
0
2
R
∗
T
P
T
_
1
η=−1
N
T
(η) δT
˜
(η) dη
The relation between the iterative traction and the iterative opening represents the stiﬀness
of the cohesive zone.
δT
˜
(η) =
∂T
˜
∂∆
˜
δ∆
˜
=
_
¸
¸
_
∂T
t
∂∆
t
∂T
t
∂∆
n
∂T
n
∂∆
t
∂T
n
∂∆
n
_
¸
¸
_
_
δ∆
t
δ∆
n
_
= M(η) δ∆
˜
(η)
The iterative opening is now interpolated analoguously to the interpolation of the weighting
function. Also the rotation of the local to the global coordinate system is applied. The
iterative opening is expressed in the iterative nodal displacements in the global directions.
The iterative tractions are then expressed in the iterative nodal displacements. The iterative
internal load column can now be expressed in the iterative opening, using the stiﬀness matrix
K
∗
cz
of the cohesive zone element.
163
Examples
The cohesive zone element is placed between two bulk elements, which have linear elastic
material behavior. The ﬁgure below shows the opening and sliding mode loading, where the
displacements of the upper nodes are prescribed. The plots show the reacton forces in x−
and y−direction in the upper nodes. The forces are the same as the traction opening relaton
of the cohesive zone.
1 2
3 4
5 6
7 8
−1 0 1 2 3 4
−1
−0.5
0
0.5
1
np7
np8
f
y
/
T
n
,
m
a
x
uy/δn
1 2
3 4 5 6
7 8
−3 −2 −1 0 1 2 3
−1
−0.5
0
0.5
1
np7
np8
f
x
/
T
t
,
m
a
x
ux/δt
Fig. 14.13 : Normal and tangential opening of a cohesive zone element.
In the next example the cohesive zone is opened in normal and tangential direction simul
taneously without rotation. The plots show reaction forces in the upper nodes and the
tractionopening behavior in the integration points of the cohesive zone element. Due to the
coupling in the tractionopening law, the normal and tangential tractions do not reach their
maximum values.
1 2
3 4
5 6
7 8
0 1 2 3 4 5
−0.5
0
0.5
1
1.5
2
np7
np8
f
y
/
T
n
,
m
a
x
uy/δn
0 0.5 1 1.5 2 2.5
0
0.2
0.4
0.6
0.8
1
np7
np8
f
x
/
T
t
,
m
a
x
ux/δt
0 1 2 3 4 5
0
0.2
0.4
0.6
0.8
1
ip1
ip2
T
n
/
T
n
,
m
a
x
∆n/δn
0 0.5 1 1.5 2 2.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
ip1
ip2
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.14 : Mixed mode opening of a cohesive zone element without rotation.
The cohesive zone is now opened in normal direction with a diﬀerent displacement in the upper
nodes, resulting in rotation. Reaction forces in xdirection are zero. Tangential tractions in
the integration points of the czelement diﬀer also.
1 2
3 4
5
6
7
8
0 0.5 1 1.5 2 2.5
−0.2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
np7
np8
f
y
/
T
n
,
m
a
x
uy/δn
−0.5 −0.4 −0.3 −0.2 −0.1 0
−1
−0.5
0
0.5
1
np7
np8
f
x
/
T
t
,
m
a
x
ux/δt
0 0.5 1 1.5 2 2.5
0
0.2
0.4
0.6
0.8
1
ip1
ip2
T
n
/
T
n
,
m
a
x
∆n/δn
0 0.05 0.1 0.15 0.2
0
0.05
0.1
0.15
0.2
0.25
ip1
ip2
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.15 : Opening of a cohesive zone element with rotation.
164
For large rotations it is obvious that the normal and tangential directions in the cohesive zone
element are changing during the opening. In the following three analyses the reference plane
is chosen subsequently in the midplane (line AB), the bottom plane (line between node 1
and node 2) and the upper plane (line between node 4 and node 5). The choice of the local
basis has a pronounced inﬂuence on the decomposition of the opening displacements, resulting
in diﬀerent normal and tangential openings for diﬀerent local bases. The ﬁgure shows the
deformed state and the tangential traction in the cohesive zone element.
1 2
3 4
5
6
7
8
1 2
3 4
5
6
7
8
1 2
3 4
5
6
7
8
0 0.05 0.1 0.15 0.2
0
0.05
0.1
0.15
0.2
0.25
ip1
ip2
T
t
/
T
t
,
m
a
x
∆t/δt
0 0.1 0.2 0.3 0.4
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
ip1
ip2
T
t
/
T
t
,
m
a
x
∆t/δt
−0.06 −0.04 −0.02 0 0.02 0.04
−0.04
−0.03
−0.02
−0.01
0
0.01
0.02
0.03
0.04
ip1
ip2
T
t
/
T
t
,
m
a
x
∆t/δt
Fig. 14.16 : Inﬂuence of choice of reference plane.
14.4 CZ for large deformation
For large deformation it is no longer physical to discriminate between normal and tangential
openings. A large displacement formulation is therefore proposed to resolve the ambiguity
induced by the choice of a local basis, whereby no distinction will be made between normal and
tangential loadings. Instead of deﬁning two separate constitutive relations for the normal and
tangential direction, only one constitutive relation between the traction
T and the opening
displacement
∆ is used. The cohesive zone law is based on the normal traction relation of
the exponential cohesive zone law discussed before.
A unit vector with e is deﬁned between two associated material points at the interface.
The workofseparation φ is the dissipated energy after complete opening. The characteristic
opening length δ is the opening for which T reaches the maximum value T
max
.
0 1 2 3 4 5 6
0
10
20
30
40
∆ [µm]
T
[
M
P
a
]
Fig. 14.17 : Cohesive zone law for large deformations.
∆ = ∆e ;
T = T(∆)e ;
T =
f(
∆) ; T =
φ
δ
_
∆
δ
_
exp
_
−
∆
δ
_
φ =
_
∞
∆=0
T(∆) d∆ ; T
max
=
φ
δ exp(1)
165
Upon unloading, the cohesive zone law may show an irreversible response. Two types of
irreversible behavior are considered: linear elastic unloading to the origin, i.e. elasticity
based damage (i) and unloading with the initial stiﬀness of the cohesive zone, i.e. plastically
damaged (ii). Both cases are shown in the ﬁgure below.
Fig. 14.18 : Large deformation cohesive zone response for elastic and history dependent
behavior.
Modemixity
The opening mode of the cohesive zone is quantiﬁed by a modemixity parameter d. In a
twodimensional cohesive zone,
d
1
and
d
2
are the components of the normals n
1
and n
2
of the
two cohesive zone edges perpendicular to
∆. Parameter d has a value between 0 (modeI) and
2 (modeII). Intermediate values of d represent a mixedmode opening. The tractionopening
relation is extended with this modemixity d, where a parameter α controls the inﬂuence of
modemixity behavior: the interface is stronger (α > 0) or weaker (α < 0) in modeII than
in modeI.
Fig. 14.19 : Modemixity parameter for a twodimensional cohesive zone.
d = 
d
1
−
d
2

T =
φ
δ
_
∆
δ
_
exp
_
−
∆
δ
_
exp
_
α
d
2
_
166
To investigate the modemixity behavior, a single cohesive zone is opened under an angle
β and the total dissipated energy is quantiﬁed. The value of the potential was taken to be
φ = 100 Jm
−2
. The two faces are kept parallel to each other and β is deﬁned as the angle
between one of the faces and the vector
∆.
The results are shown in the ﬁgure below. Obviously, the energy dissipated in modeI
(β = 90
o
) is independent of α and equals φ. In the case of mixedmode or modeII (β = 0
o
)
opening of the cohesive zone, the total dissipated energy depends on α. In the right ﬁgure,
the traction is plotted for three values of α in modeII, where the total dissipated energy (area
under the curve) depends on the parameter α.
0 30 60 90
80
90
100
110
120
W
[
J
m
−
2
]
β [
o
]
α = 0.2
α = 0.0
α = −0.2
0 2 4 6 8 10
0
10
20
30
40
50
∆ [µm]
T
[
M
P
a
]
α = 0.2
α = 0.0
α = −0.2
Fig. 14.20 : Inﬂuence of modemixity on cohesive zone response.
14.5 Applications
Cohesive zones are used to study deadhesion between a wide range of materials under a large
scope of loading conditions. Most applications comprise small deformations and quasistatic
loading.
Large deformations occur in the next two examples : the delamination of the polymer
coating on a steel sustrate, as is used in beverage cans, and, the opening of a foodcan, where
the lid is opened by peeling oﬀ a sealing foil.
Repeated cyclic loading with small deformations can lead to fatigue damage and de
adhesion between two entities. In electronic devises, the repeated switching causes cyclic
thermal loading and leads to deadhesion between the solder and the connection pad. This
phenomenon is also studied with cohesive zone models, combined with a damage evolution
law.
14.5.1 Polymer coated steel for packaging
Traditionally, products like aerosols, beverage and food cans, beer caps and luxury products
are made from sheet metal. After the forming steps, the product is cleaned and lacquered
to prevent corrosion, to give its surface a glossy appearance and to assure good printability.
Currently, more and more products are made from polymer coated sheet metal, making
subsequent lacquering superﬂuous. This implies considerable cost savings and eliminates
the emission of volatile organic compounds. Because the coating is subjected to the same
deformation processes as the metal substrate, delamination may occur, leading to the loss of
protective and attractive properties of the product which is unacceptable. If delamination
167
can be predicted, the processing routes, parameters and tooling can be adjusted to prevent
it.
This research is described in detail in the PhDthesis of M.J. van den Bosch [8] and
papers, which can be found in the References.
Fig. 14.21 : Food and beaverage cans.
Fig. 14.22 : Deepdraw simulation and
delamination in polymersteel interface.
14.5.2 Easy peeloﬀ lid
Many food containers are sealed with a socalled Easy PeelOﬀ Lid (EPOL). The EPOL
is easy to open and to produce and consists of two parts: the protact ring and the aluﬁx.
The protact ring is made of polymer coated steel with a PolyPropylene (PP) layer at the
outside and a PolyEthyleneTerephthalate (PET) layer at the inside of the can. The aluﬁx is a
membrane, consisting of aluminum which is coated with a peelable PP layer at one side of the
foil. The thermoplastic PP layer of the protact ring and aluﬁx foil are heat sealed together,
achieving an airtight closure.
The EPOL closure must be peelable with a force not exceeding 25 N under an angle
of 135
o
, which is the industrial standard for opening peelable food containers. The EPOL
closure must also be airtight and resist a pressure buildup in the can of about 2.42.6 bar at
sterilization temperature, which also is a common requirement in the industry.
This research is described in detail in the MScthesis of S. Bijker [4].
Fig. 14.23 : Easy Peeloﬀ Lid in closed and opened state.
168
14.5.3 Solder joint fatigue
Solder joints provide electrical, thermal and mechanical continuity in electronic packages. To
day, miniaturization is the major driving force in consumer electronics design and production.
Eﬀorts in decreasing component dimensions have led to the development of ball grid array
(BGA) and ﬂip chip packages, where solder balls are employed.
These solder balls are subjected to diﬀerent types of loading:
• Thermal cycling due to repeated power switching evokes heat related phenomena: the
mismatch in the coeﬃcient of thermal expansion between the package components causes
cyclic mechanical strains.
• As a result of the multiphase nature of Sn based solder alloys and the thermal anisotropy
of βSn, internal stresses build up in the solder.
• Cyclic thermomechanical loading evokes creepfatigue damage or creep rupture.
• Bending of the board induces shear and tensile stresses in the solder joints.
The ﬁgure below shows a typical bump between two pads, where thermal loading of the
system leads to deformation and results in plastic strains. Cyclic loading leads to fatigue
damage and ﬁnally to failure. Instead of using phenomenological laws like MansonCoﬃn,
the fatigue damage is assessed with a cohesive zone approach.
This research is described in detail in the MScthesis of M.E. Erin¸ c [23] and related
papers, which can be found in the References.
Fig. 14.24 : Ball grid array (BGA) and FEmodel of one solder ball with cohesive zones.
Bibliography
[1] Barenblatt, G.I. The mathematical theory of equilibrium cracks in brittle fracture. Ad
vances in Applied Mechanics, Vol 7, 1962, pp 55129.
[2] Barsoum, R.S. On the use of isoparametric Finite Elements in Linear Fracture Mechan
ics. Int. J. Numerical Methods in Engineering, Vol 10, 1976.
[3] Bates, R.C.; Clark, W.G. Fractography and fracture mechanics. Trans. ATM, Vol 62,
1969, pp 380388.
[4] Bijker, S.Study of the opening behavior of food containers with Easy PeelOﬀ Lid. MSc
Thesis, MT 08.11, Eindhoven University of Technology, 2008.
[5] Bosch van den, M.J.; Schreurs, P.J.G.; Geers, M.G.D.An improved description of the
exponential Xu and Needleman cohesive zone law for mixedmode decohesion. Engineering
Fracture Mechanics, Vol 73, 2006, pp 12201234.
[6] Bosch van den, M.J.; Schreurs, P.J.G.; Geers, M.G.D.A cohesive zone model with a
large displacement formulation accounting for interfacial ﬁbrillation. Eur.J.Mech., Vol
26, 2007, pp 119.
[7] Bosch van den, M.J.; Schreurs, P.J.G.; Geers, M.G.D.On the development of a 3D cohe
sive zone element in the presence of large deformations. Comp.Mech., Vol 42, 2008, pp
171180.
[8] Bosch van den, M.J.Interfacial delamination in polymer coated metal sheet. PhD Thesis,
Eindhoven University of Technology, 2007.
[9] Bosch van den, M.J.; Schreurs, P.J.G.; Geers, M.G.D.; Van Maris, M.P.H.F.L.Interfacial
characterization of prestrained polymer coated steel by a numericalexperimental ap
proach. Mech.Mat., Vol 40, 2008, pp 302314.
[10] Broek, D. Elementary engineering fracture mechanics; 4thed.. Martinus Nijhoﬀ Publish
ers, 1986.
[11] Broek, D. The practical use of Fracture Mechanics. Kluwer Academic Publishers, 1989.
[12] Camacho, G.T.; Ortiz, M. Computational modelling of impact damage in brittle materi
als. Int. J. of Solids and Structures, Vol 33, 1996, pp 28992938.
[13] Carlson, R.L.; Kardomateas G.A. An Introduction to Fatigue in Metals and Composites.
Chapman and Hall, 1996.
169
170
[14] Donahue, R.J.; Clark H.M.; Atanmo P.; Kumble R.; McEvily A.J. Crack opening dis
placement and the rate of fatigue crack growth. Int. J. of Fracture Mechanics, Vol 8, 1972,
pp 209219.
[15] Dugdale, D.S. Yielding of steel sheets containing slit. J. of the Mechanics and Physics of
Solids, Vol 8, 1960, pp 100108.
[16] Elber, W. Fatigue crack closure under cyclic tension. Engineering Fracture Mechanics,
Vol 2, 1970, pp 3745.
[17] Elber, W. The signiﬁcance of fatigue crack closure. Damage Tolerance in Aircraft Struc
tures, ASTM STP 486, 1971, pp 230242.
[18] Erdogan, F,; Sih, G.C. On the crack extension in plates under plane loading and trans
verse shear. J. Basic Eng., Vol 85, 1963, pp. 519527.
[19] Erdogan, F.; Ratwani M. Fatigue and fracture of cylindrical shells containing a circum
ferential crack. Int. J. of Fracture Mechanics, Vol 6, 1970, pp 379392.
[20] Erin¸ c, M.E.; Schreurs, P.J.G.; Zhang, G.Q.; Geers, M.G.D.Microstructural dam
age analysis of SnAgCu solder joints and an assessment on indentation procedures.
J.Mat.Sc.:Mat.Electr., Vol 16, 2005, pp 693700.
[21] Erin¸ c, M.; Schreurs, P.J.G.; Zhang, G.Q.; Geers, M.G.D.Characterization and fatigue
damage simulation in SAC solder joints. Microelectronics Reliability, Vol 44, 2004, pp
12871292.
[22] Erin¸ c, M.E.; Schreurs, P.J.G.; Geers, M.G.D.Integrated numericalexperimental analysis
of interfacial fatigue fracture in SnAgCu solder joints. Int.J. Solids and Structures, Vol
44, 2007, pp 56805694.
[23] Erin¸ c, M.E.Thermomechanical fatigue failure of interfaces in leadfree solders. PhD The
sis, Eindhoven University of Technology, 2007.
[24] Forman, R.G.; Kearney V.E.; Engle R.M. Numerical analysis of crack propagation in
cyclicloaded structures. J. of Basic Engineering; Trans. ASME, Vol 89, 1967, pp 459
464.
[25] Gdoutos, E.E. Fracture Mechanics ; an introduction. Kluwer Academic Publishers, 1993.
[26] Gordon, J.E. The new science of strong materials. Penguin Books, 1976.
[27] Gordon, J.E. Structures. Penguin Books, 1978.
[28] Griﬃth, A.A. Phil. Trans. Roy. Soc., Vol 221, 1921, pp 163.
[29] Hertzberg, Richard W. Deformation and fracture mechanics of engineering materials.
John Wiley and Sons, 1976.
[30] Hertzberg, Richard W. Deformation and fracture mechanics of engineering materials;
4th ed.. John Wiley and Sons, Chichester, 1996.
171
[31] Hutchinson, J.W. Singular behavior at the end of a tensile crack in a hardening material.
J. of the Mechanics and Physics of Solids, Vol 16, 1968, pp 1331.
[32] Inglis, C.E. Stresses in a plate due to the presence of cracks and sharp corners. Transac
tions of the Institute of Naval Architects, Vol 55, 1913, pp 219241.
[33] Irwin, G.R.; Kies, J.A. Welding J. Res. Suppl., Vol 33, 1954, pp 1935.
[34] Irwin, G.R.; Kies, J.A.; Smith, H.L. Fracture strength relative to onset and arrest of
crack propagation. Proceedings of the American Society for Testing Materials, Vol 58,
1958, pp 640657.
[35] Irwin, G.R.; Handbook of physics. Vol 6, 1958, pp 551590.
[36] Irwin, G.R. Plastic zone near a crack and fracture toughness. Proc. 7th Sagamore Conf.,
1960, pp IV63.
[37] Johnson, Gordon R.; Cook, William H. Fracture characteristics of three metals subjected
to various strains, strain rates, temperatures and pressures. Engineering Fracture Me
chanics, Vol 21, 1985, pp 3148.
[38] Kanchanomai, C.; et al. Low cycle fatigue test for solders using noncontact digital image
measurement system. Int.J. of fatigue, Vol 24, 2002, pp 5767.
[39] Kanninen, M.F.; Popelar, C.H. Advanced fracture mechanics. Oxford University TTss,
New York, 1985.
[40] Klingbeil, N.W. A total dissipated energy theory of fatigue crack growth in ductile solids.,
2003.
[41] Knott, J.F. Fundamentals of fracture mechanics. Butterworths, 1973.
[42] Kurz, Wolfgang Debonding along a ﬁber/matrix interface in a composite. WFW 93.109,
1993(jul), pp 34.
[43] Lim, I.L.; Johnston, I.W.; Choi, S.K. Application of singular quadratic distorted isopara
metric elements in linear fracture mechanics. Int. J. for Numerical Methods in Engineer
ing, Vol 36, 1993, pp 24732499.
[44] Matous, K.; Geubelle, P. Multiscale modeling of particle debonding in reinforced elas
tomers subjected to ﬁnite deformations. International Journal for Numerical Methods in
Engineering, Vol 65, 2006, pp 190223.
[45] McMaster (Ed.). Nondestructive Testing Handbook....
[46] Mediavilla, J. Continuous and discontinuous modelling of ductile fracture. PhD Thesis
TU/e, 2005.
[47] Miner, M.A. Cumulative damage in fatigue. Journal of Applied Mechanics, Vol 12, 1945,
pp A159A164.
[48] Murakami, Y. Stress intensity factors handbook, Vol.1 and Vol.2. Pergamon, N.Y., 1987.
172
[49] Murakami, Y. Stress intensity factors handbook, Vol.3. Pergamon, N.Y., 1992.
[50] Neuber, H. Theory of stress concentrations for shearstrained prismatical bodies with
arbitrary nonlinear stressstrain law. J. Appl. Mech. ASME Trans., Vol 28, 1961, pp
544550.
[51] Paris, P.; Erdogan, F. A critical analysis of crack propagation laws. J. of Basic Engineer
ing; Trans. ASME, Vol 85, 1963, pp 528534.
[52] Peterson, R.E. Stress concentration design factors. John Wiley, 1953.
[53] Peterson, R.E. Stress concentration factors. Wiley, New York, 1974.
[54] Rice, J.R. A path independent integral and the approximate analysis of strain concentra
tions by notches and cracks. J. of Applied Mechanics, Vol 35, 1968, pp 379386.
[55] Rice, J.R.; Rosengren, G.F. Plane strain deformation near a crack tip in a powerlaw
hardening material. J. of the Mechanics and Physics of Solids, Vol 16, 1968, pp 112.
[56] Roberts, D.K.; Wells, A.A. The velocity of brittle fracture. Engineering, Vol 171, 1954,
pp 820821.
[57] Rooke, D.P.; Cartwright, D.J. Compendium of stress intensity factors. Her Majesty’s
Stationary Oﬃce, London, 1976.
[58] Schijve, J. Some formulas for the crack opening stress level. Engineering Fracture Me
chanics, Vol 14, 1981, pp 461465.
[59] Sih, G.C.; Chen, E.P. Fracture analysis of unidirectional composites. J. of composite
materials, Vol 7, 1973, pp 230244.
[60] Sih, G.C. Handbook of stress intensity factors. Lehig University, Bethlehem, Pennsylvania
18055, 1973.
[61] Sih, G.C. Strain energy density factor applied to mixed mode crack problems. Int. J. of
Fracture, Vol 10, 1974, pp 305322.
[62] Tada, H.; Paris, P.C.; Irwin, G.R. The stress analysis of cracks handbook. Del Research,
1973.
[63] Tada, H.; Paris, P.C.; Irwin, G.R. The stress analysis of cracks handbook; 2nd Ed. Paris
Productions, Inc., St. Louis, 1985.
[64] Tvergaard, V. Eﬀect of ﬁbre debonding in a whiskerreinforced metal. Materials Science
and Engineering A, Vol 125, 1990, pp 203213.
[65] Tvergaard, V.; Hutchinson, J.W. The relation between crack growth resistance and frac
ture process parameters in elasticplastic solids. J. Mech. Phys. Solids, Vol 40, 1992, pp
13771397.
[66] Vinals, J. High performance fracture approach to fatigue crack analysis and life prediction.
In: WCCM5, 2002, 2002.
173
[67] Wasiluk, B.; Hoshide, T. The fracture process in elasticplastic materials under biaxial
cyclic loading., 2003.
[68] Wells, A.A. Application of fracture mechanics at and beyond general yielding. British
Welding Research Ass. Rep., MI3, 1963.
[69] Xu, X.P.; Needleman, A. Void nucleation by inclusions debonding in a crystal matrix.
Model Simul Mater Sci Engng, VOL 1, 1993, pp 111132.
[70] Yoﬀe, E.H. The moving Griﬃth crack. Philosophical Magazine, Vol 42, 1951, pp 739750.
External sources of ﬁgures and other data
Source
ﬁg 1.7 : Copied from Internet
ﬁg 1.8 : Copied from Internet (MaTe website)
ﬁg 1.10 : Copied from Kanninen [39]
ﬁg 1.12 : Unknown source
ﬁg 2.1 : Copied from Hertzberg [30]
ﬁg 2.3 : Copied from Hertzberg [30]
ﬁg 2.4 : Copied from Internet : Google Books : Derek Hull
”Fractography: observing, measuring, and interpreting
fracture surface topography”
ﬁg 3.1 : Copied from Internet : www.ventioelde.de (2009)
ﬁg 3.2 : Copied from Internet : Univ. of Strathclyde (2006)
ﬁg 8.3 : Copied from Kanninen [39]
ﬁg 9.5 : Copied from Hertzberg [30], Gdoutos [25], Kanninen [39]
ﬁg 11.16 : Copied from Kurz [42]
ﬁg 11.17 : Copied from Mediavilla [46]
ﬁg 12.1 : Copied from Hertzberg [30]
ﬁg 12.5 ; Unknown source
ﬁg 12.7 ; Unknown source
ﬁg 12.23 : Unknown source
page 139 : Data taken from Hertzberg [29]
page 139 : Figures copied from Hertzberg [29]
Appendix A
Laplace equation
The derivatives of the complex function f w.r.t. x
1
and x
2
are calculated ﬁrst. After using
tha CauchyRiemann equations and a second derivation w.r.t. x
1
and x
2
, it follows that the
real part of the complex function f satisﬁes the Laplace equations. The same can be deribed
for the imaginary part of f.
∂f
∂x
1
=
∂f
∂z
∂z
∂x
1
=
∂f
∂z
= f
′
(z)
∂f
∂x
2
=
∂f
∂z
∂z
∂x
2
= i
∂f
∂z
= if
′
(z)
_
¸
¸
_
¸
¸
_
→ f
′
(z) =
∂f
∂x
1
= −i
∂f
∂x
2
→
∂φ
∂x
1
+i
∂ζ
∂x
1
= −i
_
∂φ
∂x
2
+i
∂ζ
∂x
2
_
=
∂ζ
∂x
2
−i
∂φ
∂x
2
CauchyRiemann equations
∂φ
∂x
1
=
∂ζ
∂x
2
and
∂ζ
∂x
1
= −
∂φ
∂x
2
→
∂φ
∂x
2
= −
∂ζ
∂x
1
→
∂
2
φ
∂x
2
1
=
∂
2
ζ
∂x
1
∂x
2
;
∂
2
φ
∂x
2
2
= −
∂
2
ζ
∂x
1
∂x
2
→
∂
2
φ
∂x
2
1
+
∂
2
φ
∂x
2
2
= 0
a1
a2
Appendix B
Derivatives of Airy function
The general solution of the complex biharmonic equation is an expression for the Airy func
tion in terms of complex functions.
ψ =
1
2
_
¯ zΩ +z
¯
Ω +ω + ¯ ω
¸
The derivatives w.r.t. the complex variables z and ¯ z can be determined.
∂ψ
∂z
=
1
2
_
¯ zΩ
′
+
¯
Ω +ω
′
¸
∂ψ
∂¯ z
=
1
2
_
Ω +z
¯
Ω
′
+ ¯ ω
′
¸
∂
2
ψ
∂z∂¯ z
=
1
2
_
Ω
′
+
¯
Ω
′
¸
∂
2
ψ
∂z
2
=
1
2
_
¯ zΩ
′′
+ω
′′
¸
∂
2
ψ
∂¯ z
2
=
1
2
_
z
¯
Ω
′′
+ ¯ ω
′′
¸
The derivatives w.r.t. the coordinates x
1
and x
2
can now be calculated.
ψ
,1
=
∂ψ
∂z
∂z
∂x
1
+
∂ψ
∂¯ z
∂¯ z
∂x
1
=
∂ψ
∂z
+
∂ψ
∂¯ z
ψ
,11
=
∂
∂z
_
∂ψ
∂z
+
∂ψ
∂¯ z
_
∂z
∂x
1
+
∂
∂¯ z
_
∂ψ
∂z
+
∂ψ
∂¯ z
_
∂¯ z
∂x
1
= 2
∂
2
ψ
∂z∂¯ z
+
∂
2
ψ
∂z
2
+
∂
2
ψ
∂¯ z
2
ψ
,12
=
∂
∂z
_
∂ψ
∂z
+
∂ψ
∂¯ z
_
∂z
∂x
2
+
∂
∂¯ z
_
∂ψ
∂z
+
∂ψ
∂¯ z
_
∂¯ z
∂x
2
= i
_
∂
2
ψ
∂z
2
−
∂
2
ψ
∂¯ z
2
_
ψ
,2
=
∂ψ
∂z
∂z
∂x
2
+
∂ψ
∂¯ z
∂¯ z
∂x
2
=
∂ψ
∂z
i −
∂ψ
∂¯ z
i
ψ
,22
=
∂
∂z
_
∂ψ
∂z
i −
∂ψ
∂¯ z
i
_
∂z
∂x
2
+
∂
∂¯ z
_
∂ψ
∂z
i −
∂ψ
∂¯ z
i
_
∂¯ z
∂x
2
= 2
∂
2
ψ
∂z∂¯ z
−
∂
2
ψ
∂z
2
−
∂
2
ψ
∂¯ z
2
a3
a4
Contents
1 Introduction 2 Fracture mechanics 2.1 Fracture mechanisms . . . . . . . . . . . . . . . . . 2.1.1 Shearing . . . . . . . . . . . . . . . . . . . . 2.1.2 Cleavage . . . . . . . . . . . . . . . . . . . . 2.1.3 Fatigue . . . . . . . . . . . . . . . . . . . . 2.1.4 Crazing . . . . . . . . . . . . . . . . . . . . 2.1.5 Deadhesion . . . . . . . . . . . . . . . . . . 2.2 Ductile  brittle behavior . . . . . . . . . . . . . . . 2.2.1 Charpy vnotch test . . . . . . . . . . . . . 2.3 Theoretical strength . . . . . . . . . . . . . . . . . 2.3.1 Discrepancy with experimental observations 2.3.2 Griﬃth’s experiments . . . . . . . . . . . . 2.3.3 Crack loading modes . . . . . . . . . . . . . 3 Experimental techniques 3.1 Surface cracks . . . . . 3.2 Electrical resistance . 3.3 Xray . . . . . . . . . 3.4 Ultrasound . . . . . . 3.5 Acoustic emission . . . 3.6 Adhesion tests . . . . 1 9 9 9 10 11 12 12 13 13 15 16 17 18 19 19 20 20 20 21 21 23 23 24 25 26 27 28 28 28 29
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
4 Fracture energy 4.1 Energy balance . . . . . 4.2 Griﬃth’s energy balance 4.3 Griﬃth stress . . . . . . 4.3.1 Discrepancy with 4.4 Compliance change . . . 4.4.1 Fixed grips . . . 4.4.2 Constant load . . 4.4.3 Experiment . . . 4.4.4 Examples . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . experimental observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . 6. . . . . .2 Displacement . . . . . . . . . . . . . .3 Linear elastic material behavior 5. . . . . . . . . .4 Equilibrium equations . . . . 6. . . . . . . .5. 6. . . . . . . . . . . . .4 Stress intensity factor . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . .2 Displacement . . . . .4 Mode II . . . . . . . .9 Circular hole in ’inﬁnite’ plate . . . 6. . . . . . .1. . . . . . . . . . . . . .3 Choice of complex functions 6. . . . . . . 5. . . .5. . .2 Stress components . . . . . . . . . . . . .2 Stress intensity factor .2. . . . . . . . . . . . . . . .1. . . . . . . . . . . . . .1 KIc values . 6. . . . . . . . . . . . . . . . . . . .3 Crack tip solution . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . .1 Complex variables . . . . . . . .4 Displacement components . . . . . . . . . . . . . . . . . . . . .4. . .5 Mode III . . . . . . . . . . . .3. . . . 5.4 Crack tip solution . . .7 Displacement method . . . .5. . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . 5. . . . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . 31 31 33 33 34 35 36 37 37 39 44 45 45 45 46 46 47 47 48 48 50 50 51 51 51 53 53 53 53 54 54 54 55 55 55 56 56 56 57 57 59 59 61 62 . . . .1 Kzone . . . . . . . . . . . . . . . . .9 Relation G − K .10. . . . . . . . . . . . . . . . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . .6 Crack tip stress (mode I. . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . II. . . . 6. . . . . . . . . . . . . . . . . 6. . . . . . . . .2. . . 6. . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . 6. .1. . . . .II 5 Stress concentrations 5. . . . .3.8 Kbased crack growth criteria . . . . . . . . . . . . . . . . . . . . . . . .4. . . . . . . . . . .5 Crack tip solution . . . . .3 Stress components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . .3 Mode I . . 6. . . . . . . . .6 Plane strain . . . . . . . . . .2 Solution of biharmonic equation . .10 The critical SIF value . . . . . . . . . . . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . 5. . . . . . . . . 5. 6. . . .3 Stress intensity factor . . . . . . . . . . . . . . . . . . . . . . . . . .1 Laplace equation . . . . . . . . . . . . .1 Stresses . . . . . . . . . .1 Displacement . . . . . . . . . . . . . . .2 Stress . . . . . . . .8 Stress function method . . . . . . . 6. .2. . . . . 6. . .1 Complex plane . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Deformation and strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. . . . . . . . . . . . . .4 Biharmonic equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Elliptical hole . . . . . . .3 Laplace operator . . . . . . . . . . . . 6. . . . . . . . . . . . . . . . . . . . . .2 Complex functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. 6. . . . . . . . . . . . . . . . . . . . . . .5. . . . . . 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . .5 Plane stress .6. 5. . . . . .7 SIF for speciﬁed cases . 6. . . . . . . . . . . . . . 5. . . . . . . . . . . . . . . . III) . . . . . . . . . . . . . . . . . . . . .1 Displacement . . . . . . . . . . . . 6. . . . . . . . . . 6. . . . . . . . . . . . 6 Crack tip stresses 6. . . . . . . . . .
. . . . . . . . . . . . . . . 8. . . . .3 Plastic zones in the crack plane .1 Von Mises and Tresca yield criteria . . . 9. . . . . . . . .2 Multimode load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9. . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . 9 Plastic crack tip zone 9. . . . . . . . . . . . . . .3 Relation J ∼ K . . . . . . .8.3. . . . . . . . . . . . . . . . . . . . . . . . . . .5 Inﬂuence of the plate thickness . . . . . 7. . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8. . . .7 Plastic constraint factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . . . . . . . . . . . .1 RambergOsgood material law . . . . . . . . . . . . . . . . . . . . . . . .1 Irwin plastic zone correction . . . . . . . . . . . . . . . . . . . . . . . . . .2 DugdaleBarenblatt plastic zone correction 9. . . . . . . .3 Special elements . . . .3 Von Mises plastic zone .8. . . .2 Fast fracture and crack arrest 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 . . . . . . . . . . . . . . . . . .2 Jintegral . . .3. . .4 Tresca plastic zone . . . . .III 7 Multimode crack loading 7.2 Strain energy density (SED) criterion 8 Dynamic fracture mechanics 8. . . .2 Path independence . . . . . . . . . . . . . 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Integral along closed curve . . . 11 Numerical fracture mechanics 105 11. . . . . . . . . . . . . .1 Cracktip opening displacement . . . . 10. . . . . . . . 106 11. . . . .2. . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . .3 HRR crack tip stresses and strains . . . . . . . . . . . . . . .2. . . . . . . . . . . . . . . . .3. . 10. . . . . . . . . . . . . . . . . 8.2 Principal stresses at the crack tip . . 10. . . .2 CTOD by Dugdale . . . . . . . . . . . . .1 Quadratic elements . . . . . . .1 Crack growth rate . . . 7. . . . .9 Small Scale Yielding . . . . . . . . . . .1 Maximum tangential stress criterion . . . . . . . . . . . . . . . . . . . . . . . . . .8 Plastic zone in the crack plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 CTOD crack growth criterion . . . . . . 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . . . . . . . . . . . . . . . . . . . . .1 CTOD by Irwin . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . 9. . 63 63 67 69 69 72 77 77 80 80 81 81 82 83 83 84 85 86 87 88 89 90 90 91 92 92 95 95 96 97 98 98 99 99 100 101 102 102 103 .2 Crack tip mesh . . . . . . . . 7. . . . . . . . . . . . . .1 Crack branching . . . . . . .2 Elastic wave speeds . . . . . . . .3. . 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10. . . . . 7.3 Crack growth direction . . . . .4 Experiments . 10. . . . . . . . . .3 Jintegral crack growth criterion .8. . .1. . . . . . . . . . 10. . . . . . 10. . . . . . . . . . . . . . . . . . . . .1. . 9. . 10. 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Nonlinear Fracture Mechanics 10. . . . . . . . . . . . . . . . . . . . . 9. 9. .3 Crack tip stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . .1 Stress component transformation .6 Shear planes . . . . . . . . . . . . . . . . . . . 9.2 HRRsolution . . . . . 10. . . . .
. . . . . . . . . . . 12. . . . . . . . . . . . . . 12. . . . . . . . 12. . . . . .4. . .1. . . . . . . . . . . . .8. . . .2 Properties of engineering plastics . . . . . . . . . . . . . . . . . . . . . . . . . .1 Fatigue limit . . . . . . .1 Damage . . . . . . . . . . . . . . . . . 144 . . . . . . . 12. . . . . . 12. . . . . . . . . . . . . .5. . . . . . . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. . . . . . .5 High/low cycle fatigue . . . . . . . . . . . . . .1 Crack growth models . . . . . fatigue .1 Mechanical properties . . . . . . . .4 (PSN)curve . . .8. . . . . .4 Material properties . . . . . . . . . . . . . . . . . . . . . .1 Domain integration . . . . . . . .8. . . . . . . . . . . . . . . 11. . . . 11. . .7. . . . . . . . . 12. . . . . . . . . . . 12. . 12. . . . . . . . . . . .3. . .3 Inﬂuence of average stress 12. . . . . . . . 12. . . . . .4. . . . . . . . . . . . . . . . . . . . . . . .5 Surface quality . . . . . . .1 Random load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . .6 Load spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. .7 Jintegral . . . . .3. . . .4. . . .3. . . . . . . . .8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4. . . . 11. . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . .4. . . . . . .2 Moving Crack Tip Mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . . . . . . . . . 12.5. . . 144 13. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. . 12. . . .4 Inﬂuence factors .4. . . . . 11. . . . . . . . 12. . . . . . 12 Fatigue 12. . . . . . . . . . . . . . . . . . .2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Tensile overload . . . . . . . . . . . . . . . .4 Other crack grow laws . . . . . . . . . . . .1 Load spectrum . . . . . . . . . . . . . . . . . . . . . . . . 11. . . 12. . . . . . . . .2 Paris law . . . .3. . . . . . . . . . . . . . . . . . .3. . . . . . .4 Quarter point elements . . . . . . . . . . . . . . . . 12.3. . . . . . . . . . . . . . . . . . 12. . . . . . . . . . . . .3 Stress gradients . . . . . . . . . .7 MansonCoﬃn relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. . . . . . . . 12. . . . . . . . . . . . . . . . . . . . . . . . 11. . . . . . .6. . 11. . . . . . 107 108 110 111 112 112 113 114 114 114 115 116 119 119 120 120 121 121 122 123 124 124 125 125 126 126 126 127 127 128 128 128 129 130 132 133 134 135 136 138 141 . . . . . 143 13. . . . . . . . 12. . . . . . . . . . . . . . .6 Stress intensity factor . . . . .4 Smeared crack approach . 12. . . . . . 11. . . . . . . . . . . . . . . . . . . . . . . . . . .1 Node release .3 Fatigue load . . . . . . . .6 Basquin relation . . . . . .4. . . . . . . . . .5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. . . .3. . . . . . . . . . .8 Total strainlife curve . . . . . .1 Onedimensional case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Fatigue life . . . . . . . . . . . . . . . . . . . . . . .2 Stress concentrations . . .2 De Lorenzi Jintegral : VCE technique . 12. . . .2 (SN)curve . . . . . . . . . 12. . . . . . . . . . . . . . . . . . . . .8 Crack growth simulation . . . . . . . . . . . . . . . . 12. . . . 12. . .1 Crack surface . . . . . . . . . .3 Element splitting . . . . . . . . . . . . . . . . . .7 Design against fatigue .6 Environment . . . .5 Crack growth . 13 Engineering plastics (polymers) 143 13. . . . . . . . . . . . . . . . . .5.5 Virtual crack extension method (VCEM) . . . . . . . . . . . . .IV 11. . . . . . . . . . . . . . . . . . . . . . . .5 Crack growth at low cycle 12. . . . . . . . . . . . . . . . . . . . . .6. .
. . . . .3 Rigidlinear . . . . . . . . .1. . . . . . A Laplace equation B Derivatives of Airy function . . . . . . . . . . . . . . . . . . . .1 Local vector base . . . . . . . . . . . . . . . .2 Piecewise linear . . . . . . .5. . . .3 Solder joint fatigue . . . . . . . . . . . . . . . . . .2 Weighted residual formulation with cohesive 14. . . . .3. .2 Fatigue . . . . .1 Cohesive zone models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Exponential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Twodimensional CZ element . . . .5 Applications . . . . . . . . . . . . . . . .5. . . . 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. . . . . 14. . . 14. . . . . . . . . .1. . .4 CZ for large deformation . . . . . . . . . . . . . . . . . .2 Easy peeloﬀ lid . . 14. . . . . . . . . . . . . . . 14. . . . . . . . . .1 Polynomial . . . . . 14. . .5. .1 Polymer coated steel for packaging . . . . . . . . . . . . . . . . . . . . . . .1. . . . . . . . . . .V 13. . . . . . . . . 14. . . . . 145 149 149 150 151 151 152 157 160 160 162 164 166 166 167 168 a1 a3 14 Cohesive zones 14. . . . . . zones . . . . . . . . . . . . . . . . .1. 14. . . . . . . . . . . . . . . . . . . .3. . . . . 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14. . . . . . . . .2 Iterative procedure . . . 14. . . . . . . . . . . . . . .
.
e. being the resistance against reversible deformation. although this is associated with structural changes. The only requirement is that the material behavior is studied on a scale.Chapter 1 Introduction Important aspects of technological and biological structures are stiﬀness and strength. is always required to be high. molecular slip and breaking of atomic bonds. may vary over a wide range.g. the resistance against irreversible deformation. Requirements on stiﬀness. 1. phase transformation. Continuum mechanics When material properties and associated mechanical variables can be assumed to be continuous functions of spatial coordinates. 1 . large enough to allow small scale discontinuities to be averaged out. Strength.1 : Stiﬀness and strength. because this deformation may lead to loss of functionality and even global failure. Fig. dislocation movement. analysis of mechanical behavior can be done with Continuum Mechanics. This may also apply to permanent deformation.
2 : Deformation of a continuum. 3). 2.j + ρqi = 0 ρ qi pi = σij nj σij = Nij (εkl ) . 2. 2. volume / area base vectors position vector displacement vector strains compatibility relations equilibrium equations density load/mass boundary conditions material model V0 . 3) by diﬀerentiation – ( ).k ) σij. The strain components. model and characterize the material behaviour. The stress components σij (i. σij = σji Material behavior To investigate. x u 1 εkl = 2 (uk. V0 A0 V u x0 e3 A x e2 O e1 Fig. j = 1. e2 . which is characterized by a material model. it is necessary to do experiments.l + ul. 3) must satisfy the equilibrium equations – partial diﬀerential equations – and boundary conditions. A {e1 .2 When a threedimensional continuum is subjected to external loads it will deform. 3). 1. relating stress components σij to strain components εkl (k. V / A0 . which are derived from the displacements ui (i = 1. 2. l = 1. The most simple experiment is the tensile/compression loading of a tensile bar. . e3 } x0 .j – with respect to spatial coordinates xj (j = 1. In most cases the equilibrium equations are impossible to solve without taking into account the material behavior. are related by the compatibility relations.
timedependent. elastoplastic : irreversible. the averaging procedure can no longer be applied and discontinuities must be taken into account. σ σ σ ε ε σ ε ε Fig. damage : irreversible. viscoplastic. viscoplastic : irreversible. This localization results in a macroscopic crack.5 : Stressstrain curves for hardening (left) and sofening (right) material behavior. resulting in global failure. timedependent. which can be concluded to be elastic : reversible.4 : Stress excitation and strain response as function of time. decreasing properties. . elastoplastic. timeindependent. hardening and softening behavior.3 : Tensile test. 1. 1. Fracture When material damage like microcracks and voids grow in size and become localized. 1. viscoelastic. viscoelastic : reversible. Stressstrain curves may show linear. Topleft →anticlockwise: elastic. timeindependent. σ t1 t2 t ε t1 t2 t σ t1 εe εp t1 t2 t t2 t ε t1 t2 t σ t1 t2 t ε σ t1 t2 t ε t1 t2 t Fig.3 Straintime and stresstime curves reveal much information about the material behavior. Fig. which may grow very fast. nonlinear.
theory and terminology.6 : Tensile test with axial elongation and fracture. experimental and theoretical mechanics.4 Although early approaches have striven to predict fracture by analyzing the behavior of atomic bonds. which allow answers to questions like: • Will a crack grow under the given load ? • When a crack grows. Theoretical concepts and experimental techniques have been and are being developed. Fig. each with its own concepts. 1. Fracture mechanics In fracture mechanics attention is basically focused on a single crack. theoretical and numerical mathematics. the ﬁeld of fracture mechanics can be subdivided in several specializations. Griﬃth has shown in 1921 that attention should be given to the behavior of an existing crack. . As a result. what is its speed and direction ? • Will crack growth stop ? • What is the residual strength of a construction (part) as a function of the (initial) crack length and the load ? • What is the proper inspection frequency ? • When must the part be repaired or replaced ? Several ﬁelds of science are involved in answering these questions : material science and chemistry.
Fig. The Griﬃth criterion states that ”crack growth will occur. The . (Source: internet) Experimental fracture mechanics Detection of cracks is done by experimental techniques. Experimental Fracture Mechanics (EFM) is about the use and development of hardware and procedures.8 : Experimental tensile equipment. moreover. This is the case for Linear Elastic Fracture Mechanics (LEFM). 1. (Source: Internet) Linear elastic fracture mechanics A large ﬁeld of fracture mechanics uses concepts and theories in which linear elastic material behavior is an essential assumption. for the accurate determination of its geometry and loading conditions.5 Fig. but. when there is enough energy available to generate new crack surface.” The energy release rate is an essential quantity in energy balance criteria. 1. not only for crack detection. ranging from simple and cheap to sophisticated and expensive. Prediction of crack growth can be based on an energy balance.7 : Crack in a bicycle crank.
It is characterized by the stress intensity factor. Theories and methods for this purpose are the subject of Dynamic Fracture Mechanics (DFM). .6 resulting crack growth criterion is referred to as being global. It is also essential to predict the speed and direction of its growth. The resulting crack growth criterion is referred to as local. because a rather large volume of material is considered. 1. Using yield criteria (Von Mises. This stress ﬁeld can be determined analytically. In reality this is obviously not possible: plastic deformation will occur in the crack tip region. Fig. LEFM concepts can be used. When this zone is small enough (Small Scale Yielding. because attention is focused at a small material volume at the crack tip. the crack tip plastic zone can be determined. 1. Assumption of linear elastic material behavior leads to inﬁnite stresses at the crack tip. SSY).10 : Impact of a bullet in a plate [39]. Dynamic fracture mechanics It is important to predict whether a crack will grow or not. Fig. The crack growth criterion can also be based on the stress state at the crack tip. Tresca).9 : Plastic strain zone at the crack tip.
Crack growth criteria can no longer be formulated with the stress intensity factor. For more practical cases numerical techniques are needed. in polymers and composites). 1. based on the Crack Tip Opening Displacement. (A quarter of the plate is modelled and shown. Numerical calculations are mostly done using the Finite Element Method (FEM). This is also the case when the material behavior is nonlinear elastic (e. the crack will gradually grow. be it harmonic or random. is the Jintegral. which characterizes the stress/deformation state in the crack tip zone.g. Fig. although the load amplitude is very small.11 : Loading and deformation of a cracked plate. This phenomenon is called Fatigue Crack Propagation (FCP) and may unexpectedly and suddenly result in Fatigue Failure. Another crack growth parameter. Special quarterpoint crack tip elements must be used to get accurate results. In ElasticPlastic Fracture Mechanics (EPFM) or NonLinear Fracture Mechanics (NLFM) criteria are derived. much used in NLFM. the stress and strain ﬁelds from LEFM are not valid any more. Its calculation is possible using models of Irwin or DugdaleBarenblatt for the crack tip zone.) Fatigue When a crack is subjected to a timedependent load. Numerical techniques Analytical calculation of relevant quantities is only possible for some very simple cases.7 Nonlinear fracture mechanics When the plastic crack tip zone is too large. .
with an emphasis on a clear introduction of concepts and derivation of variables. the reader is referred to the specialized publications. For more detail. a lot of which are rather intimidating for a ﬁrst reader. Outline Many books and papers on fracture mechanics and fatigue exist. because they are either too specialized and/or lack a proper introduction of concepts and variables. The main goal is to give an overview of the whole ﬁeld.12 : Engine and crank axis. 1.8 Fig. some of which can be found in the References. This is partly due to the fact that Fracture Mechanics has developed into many specialized subjects. all focused on diﬀerent applications. . In this course a number of these subjects are discussed.
1. The dislocations will coalesce at grain boundaries and accumulate to make a void. The next mechanisms are generally distinguished. Fracture Mechanics is thus started with the experimental studies by Griﬃth in 1921 and has since then developed in various directions.Chapter 2 Fracture mechanics In this chapter we ﬁrst discuss some mechanisms. Plastic deformation is essential. this mechanism is referred to as shearing. 9 . Because the internal structure is changed irreversibly.1 Fracture mechanisms The way a crack propagates through the material is indicative for the fracture mechanism. Also the number of dislocations will increase. the macroscopic deformation is permanent (plastic). so this mechanism will generally be observed in FCC crystals. which have many closedpacked planes. They are generally related to the amount of dissipated energy during crack growth and therefore to ductile or brittle fracture. The result of these very approximating calculations appears to be not in accordance with experimental ﬁndings. which can be recognized after fracture has occured. dislocations will start to move through the lattice due to local shear stresses. The theoretical strength of a material can be predicted from the maximum bond strength between atoms. Visual inspection of the fracture surface gives already valuable information. Because the origin and growth of cracks is provoked by shear stresses. These voids will grow and one or more of them will transfer in a macroscopic crack. 2.1 Shearing When a crystalline material is loaded. One or more cracks may then grow and lead to failure. • shear fracture cleavage fracture fatigue fracture crazing deadhesion • • • • 2. Griﬃth was the ﬁrst to recognize that attention must be focused to imperfections like already existing cracks.
For some illustrative pictures. 2. we refer to the book of Broek [10]. It will also be observed when plastic deformation is . having HCP or BCC crystal structure.2 Cleavage When plastic deformation at the crack tip is prohibited.or transgranular cleavage.1.10 The fracture surface has a ’doughlike’ structure with dimples. This cleavage fracture will prevail in materials with little or no closedpacked planes. When the crack propagates along grain boundaries. resulting in dimples in the crack surface [30]. it is referred to as intergranular cleavage. 2. Fig.1 : Dislocation movement and coalescence into grain boundary voids. the crack can travel through grains by splitting atom bonds in lattice planes. the shape of which indicate the loading of the crack. This is called intra.
Fig. which mark the locations of the crack tip after each individual loading cycle. . provided that the stress is high enough. As will be described later.3 Fatigue When a crack is subjected to cyclic loading. 2. a threedimensional stress state may also result in this mechanism. The discontinuity of the lattice orientations in neighboring grains will lead to socalled cleavage steps. Intergranular cleavage will be found in materials with weak or damaged grain boundaries. Under a microscope ’striations’ can be seen. but not too high to cause sudden global fracture. Nice pictures can be found in [11]. The 1 number of cycles to failure Nf is strongly related to the stress amplitude ∆σ = 2 (σmax −σmin ) 1 and the average stress σm = 2 (σmax + σmin ). This mechanism is referred to as fatigue. The crack surface has a ’shiny’ appearance.3 : Clam shell fatigue crack surface and striations [30]. Nice pictures of macroscopic and microscopic fatigue crack surfaces can again be found in [11]. The latter can be caused by environmental inﬂuences like hydrogen or high temperature. With the naked eye we can see a ’clam shell’ structure in the crack surface. 2.11 prohibited due to low temperature or high strain rate.1. the crack tip will travel a very short distance in each loading cycle. Because crack propagation is very small in each individual load cycle.and intragranular cleavage fracture. which resemble a ’river pattern’. 2.2 : Inter. intragranulair intergranulair Fig. a large number of cycles is needed before total failure occurs.
Fig.12 2. due to solidiﬁcation and cooling . glass – and organic. Some nice pictures of crazes can be found in [29]. The adhesion of the layer to the substrate is determined by chemical bond strength and highly inﬂuenced by initial stresses and damage in the surface layer and the roughness of the substrate.1. one or a limited number of these crazes grow locally to generate a large and fatal crack. Surface layers may be metallic – metals (M) and their nitrides (MNx) or oxides (MOx) –. diamond – or organic – polymers. This is indicated as ”stress whitening”.1.4 : Crazes in polystyrene on a macroscopic and microscopic scale. (Source: Internet) 2. Substrates can also be metallic – metals and alloys –. carbides. while cohesion refers to bonding between atoms of one and the same material. In other circumstances the crazes spread out over a larger area. 2. resulting in a white colored appearance. The adhesion strength can be characterized by the maximum normal strength. anorganic – ceramics. Initial residual stresses can be classiﬁed as thermal .4 Crazing In a polymer material submicrometer voids may initiate when a critical load level is exceeded. Sometimes. the maximum .5 Deadhesion Adhesion refers to bonding between atoms of diﬀerent materials. due to volume mismatch (penetration) and plastic strain (impact). Thicknesses vary from several nanometers to hundreds of micrometers. anorganic – oxides. Applications include the joining of two diﬀerent entities such as structural parts and the adhesion of a generally thin layer on a generally thicker substrate. thin ﬁlm or coating. because the crazes refract the light. and mechanical . In the latter case the thin layer is also often referred to as surface layer.
2 Ductile . PTFE 0 10 100 ε (%) Fig. Although boundary and environmental conditions are of utmost importance. Modeling of the adhesion strength is often done with cohesive zone (CZ) models. The specimen is a beam with a 2 mm deep Vshaped notch. which has a 90o angle and a 0. is used for the generation of new crack surfaces. nylon. When a lot of available energy is used for crack growth. new free surface is generated. Not all available energy. This energy is provided by the external load and is also available as stored elastic energy.brittle behavior When a crack propagates.2. The area under the curve is a measure for the dissipated energy before failure. The main advantage of this test is that it provides a simple measure for the dissipated energy during fast crack propagation. 2.5 : Tensile test up to fracture and various stressstrain curves. mainly due to dissipative mechanisms. The load is provided by the impact of a weight at the end of a pendulum. Many experimental techniques are available to determine these parameters. which for solid materials is typically 1 [Jm−2 ]. having a speciﬁc surface energy γ. When plastic deformation and thus dissipation is less. Tensile curves for a variety of materials are shown in the ﬁgure. A crack will start at the tip of the Vnotch and runs through the specimen. A ﬁrst indication of this follows from the stressstrain curve. It is also transformed into other energies. PC PE. the standard experiment to investigate this is the Charpy Vnotch test. it is common practice to say that a certain material is brittle or ductile. The energy which is dissipated during fracture can be calculated easily . Because dissipation is associated with plastic deformation. 2. the fracture is indicated to be ductile. shear fracture is often found in materials which show ductile fracture. It is supported and loaded as in a threepoint bending test.25 mm root radius. however. The material deforms at a strain rate of typically 103 s−1 . σ ABS. where the mentioned parameters play an important role. 2.1 Charpy vnotch test Although the tensile stressstrain curve already provides an indication for brittle/ductile failure.13 tangential strength or the work needed for separation. like kinetic energy or dissipative heat. registered in a tensile test up to fracture of the tensile bar. the fracture is said to be brittle. When a lot of energy is transformed into other energies. fracture is more brittle.
For intrinsic ductile materials like FCCmetals.7Tm . More on the Charpy test can be found in [29]. the dissipated energy will be low for all T . Fig. ceramics high strength metals Al. Other tests to determine the brittleness are the Izod test and Drop Weight Test. The impact toughness can be determined for various specimen temperatures T . Ti alloys T NDT FATT Tt FTP T Fig. Zn. Cv will be high for all T .14 from the height of the pendulum weight. The dissipated energy is the Impact Toughness Cv [J].6 : Charpy Vnotch test.2Tm and for ceramics Tt = 0. For intrinsic brittle materials like high strength steel. 2. A large number of materials show a transition from brittle to ductile fracture with increasing temperature.1 ` 0.7 : Cv values as a function of temperature T . .5 ` 0. fcc (hcp) metals Cv Cv low strength bcc metals Be. 2. before and after impact. a a where Tm is the melting temperature. The transition trajectory is characterized by three temperatures : NDT FATT FTP : : : Nil Ductility Temperature Fracture Appearance Transition Temperature (Tt ) Fracture Transition Plastic As a rule of thumb we have for BCC alloys Tt = 0.
The interaction force can be approximated by a sine function.3 Theoretical strength Metal alloys consist of many crystals. The stress σ. f r σ f f x a0 1 2λ r S x Fig. with γ the speciﬁc surface energy of the material. the stored energy is released. 2.8 : Atomic bond strength between atom in a lattice. In a ﬁrst attempt to calculate the strength of such a crystalline material. can be expressed in the maximum stress σmax . It is assumed that all bonds in the area S snap at the same time and that all the stored energy is transformed in surface energy. When the interaction force and thus the interatomic distance increases. each of which has lattice planes in a certain spatial orientation. the half wavelength λ and the maximum force fmax . which. the interaction force is zero and the bond is broken. per unit of area is Ua = 2γ. The stored elastic energy per unit of area is Ui and can be calculated by integration. being the theoretical material strength. which is the ratio of this total force and the area S. The bonding force f between two atoms in two neighboring planes depends on their distance r and can be calculated from a potential. elastic energy is stored in each bond. 2 The interaction force between all atoms in an area S of the two lattice planes can be calculated by addition. one atomic plane perpendicular to the tensile load is considered. From the energy balance Ui = Ua . x = r − a0 2πx λ f (x) = σmax sin The theoretical strength σmax can be determined from an energy balance. With this approximation. available elastic energy per surfaceunity [N m−1 ] Ui = 1 S x=λ/2 f (x) dx x=0 . using the equilibrium distance a0 . when r = a0 + 1 λ. the wave length of the sine function can be calculated and the stress as a function of the displacement x can be expressed in σmax . f (x) = fmax sin σ(x) = 1 S 2πx λ . When the bond breaks at x = λ/2.15 2.
(Diamond is an exception with γ = 5 Jm−2 .1 Discrepancy with experimental observations The surface energy γ does not diﬀer much for various solid materials and approximately equals 1 Jm−2 .3. The reason for this deviation has been discovered by Griﬃth in 1921. It is clear that there is a large discrepancy between the two values : the theoretical strength is much too high. . the linear strain in the undeformed state.r.16 x=λ/2 = x=0 σmax sin λ π 2πx λ dx = σmax required surface energy [Nm−1 ] Ua = 2γ energy balance at fracture Ui = Ua → λ= 2πγ σmax [Nm−1 ] → σ = σmax sin x σmax γ The argument of the sine function is assumed to be small enough to allow a linear approximation. linearization σ = σmax sin linear strain of atomic bond x ε= a0 elastic modulus E= dσ dε = x=0 x σmax γ ≈ x 2 σ γ max εa0 2 σ γ max a0 γ → x = εa0 → σ= dσ a0 dx 2 = σmax x=0 → σmax = theoretical strength Eγ a0 σth = Eγ a0 2. is also almost the same for solids (about 10−10 m). This leads to a relation for the theoretical strength σth = σmax . The macroscopic Young’s modulus E of the material is introduced as the derivative of the stress w. The displacement x between the atomic planes is expressed in the linear strain ε.t. The table below lists values of theoretical strength and experimental fracture stress for some materials.) The equilibrium distance a0 between atoms.
4 2.3.2 6. However. 2. Griﬃth new of the earlier (1913) work of Inglis [32].17 a0 [m] E [GPa] σth [GPa] σb [MPa] σth /σb glass steel silica ﬁbers iron whiskers silicon whiskers alumina whiskers ausformed steel piano wire 3 ∗ 10−10 10−10 10−10 10−10 10−10 10−10 10−10 10−10 60 210 100 295 165 495 200 200 σth ≫ σb 14 45 31 54 41 70 45 45 170 250 25000 13000 6500 15000 3000 2750 82 180 1. March en Gordon. there was less volume and less chance for a defect to exist in the specimen.9 : Fracture strength of glass ﬁbers in relation to their thickness. Griﬃth published his work in 1921 and his paper [28] can be seen as the birth of Fracture Mechanics. σb approached the theoretical strength of 14000 MPa in the limit of zero thickness. Fracture Mechanics was born! 11000 σb [MPa] 170 10 20 d [µ] Fig. The ingenious insight that strength was highly inﬂuenced by defects has lead to the shift of attention to the behavior of cracks and the formulation of crack growth criteria.3 4. It was shown in 1976 by Parratt. who calculated stress concentrations at circular holes in plates. He reasoned that for glass ﬁbers with smaller diameters. . that surface defects instead of volume defects were the cause for the limiting strength. In the limit of zero volume there would be no defect and the theoretical strength would be found experimentally.3 4. For d > 20 µm the bulk strength of 170 MPa was found. being much higher than the nominal stress. He concluded that in his glass ﬁbers such stress concentrations probably occurred around defects and caused the discrepancy between theoretical and experimental fracture stress.7 15 16.2 Griﬃth’s experiments In 1921 Griﬃth determined experimentally the fracture stress σb of glass ﬁbers as a function of their diameter.
2.10 : Three standard loading modes of a crack. Mode I Mode II Mode III = = = opening mode sliding mode tearing mode . Mode I Mode II Mode III Fig. which are still used today [33].18 2.3 Crack loading modes Irwin was one of the ﬁrst to study the behavior of cracks.3. He introduced three diﬀerent loading modes.
ventioelde. Although it can be applied easily and onsite. Only cracks at or just below the surface can be detected in this way. 3. Experimental techniques have been and are still being developed.de (2009)) 19 . only surface cracks can be detected.Chapter 3 Experimental techniques To predict the behavior of a crack.1 Surface cracks One of the most simple techniques to reveal surface cracks is based on dye penetration into the crack due to capillary ﬂow of the dye. Some of these procedures use physical phenomena to gather information about a crack. it is essential to know its location. Magnetic ﬁelds can be visualized with magnetic particles and electric ﬁelds by the use of inertances. Experiments have to be done to reveal these data.1 : Dye penetration in the stiﬀening cone of a turbine. 3. Other techniques strive towards visualization of the crack. (Source: Internet site www. Fig. geometry and dimensions. caused by a crack. Other simple procedures are based on the observation of the disturbance of the magnetic or electric ﬁeld.
Fig. pi¨zoel. 3. crystal e wave sensor S in out ∆t t Fig.2 Electrical resistance A crack is a discontinuity in the material and as such diminishes the crosssectional area. . 3.2 : Xray robots inside and outside of a pipe searching for cracks. This may be associated with an increase of the electrical resistance. which can be measured for metallic materials and carbon composites.4 Ultrasound Visualization is also possible with sound waves.20 3.3 Xray Direct visualization of a crack can be done using electromagnetic waves. 3. (Source: Internet site University of Strathclyde (2006)) 3. Xrays are routinely used to control welds.3 : Ultrasound crack detection. This is based on the measurement of the distance over which a wave propagates from its source via the reﬂecting crack surface to a detector.
3. blade wedge test peel test (0o and 90o ) bending test scratch test indentation test laser blister test pressure blister test fatigue friction test . which can be detected at the surface. Some of them are illustrated below.6 Adhesion tests Many experimental techniques are used to determine adhesion strength of surface layers on a substrate.21 3.5 Acoustic emission The release of energy in the material due to crack generation and propagation. There is a correlation of their amplitude and frequency with failure phenomena inside the material. This acoustic emission (AE) is much used in laboratory experiments. results in sound waves (elastic stress waves).
22 .
The kinetic energy is the result of material velocity. 4.1 Energy balance Abandoning the inﬂuence of thermal eﬀects. The internally stored elastic energy is considered to be an energy source and therefore it is moved to the lefthand side of the energy balance equation. The dissipation may have various characteristics. surface energy (Ua ). It results in temperature changes. when a crack propagates. e. e. when new free surface is generated. dissipated ˙ d ) and kinetic energy (Uk ). B = thickness a A = Ba Fig. The surface energy changes. Assuming the crack to be through the thickness of a plate. the ﬁrst law of thermodynamics can be formulated for a unit material volume : the total amount of mechanical energy that is supplied to a material volume per unit of ˙ ˙ ˙ time (Ue ) must be transferred into internal energy (Ui ). ˙ energy (U The internal energy is the elastically stored energy.g.Chapter 4 Fracture energy 4. 23 . but is mostly due to friction and plastic deformation. we can also use the crack length a as a state variable.g.1 : Plate with a line crack of length a. we can also use another state variable. Instead of taking time derivatives (˙) to indicate changes. whose thickness B is uniform and constant. the crack surface area A.
so in this case of ﬁxed grips we have dUe = 0. When we gradually increase the crack length. where γ is the surface energy of the material. does not do any work. where a plate is loaded in tension and ﬁxed at its edges. energy balance energy release rate crack resistance force Griﬃth’s crack criterion dUe dUi dUa − = da da da 1 dUe dUi − G= B da da 1 dUa = 2γ R= B da G = R = 2γ [Jm−2 ] [Jm−2 ] [Jm−2 ] According to Griﬃth’s energy balance. After division by the plate thickness B. This is illustrated with the following example. due to prestraining. the increment in surface energy is higher than the available (decrease of) internal energy. The initial crack length a is indicated in the ﬁgure and it must be concluded that for crack growth da. This results in the socalled Griﬃth energy balance.24 ˙ ˙ ˙ ˙ ˙ Ue = Ui + Ua + Ud + Uk [Js−1 ] dA d d d ˙ d ()= ()=A ()=a () ˙ dt dt dA dA da dUi dUa dUd dUk dUe = + + + [Jm−1 ] da da da da da dUa dUd dUk dUe dUi − = + + da da da da da [Jm−1 ] 4. where elastic energy is released. the crack becomes unstable at the length ac . The surface energy. which equals 2γ. Dissipation and kinetic energy are neglected. . The increase of surface energy equals the decrease of the internal energy. which is associated with the crack surface is 2γa. Because the edges are clamped. when the energy release rate equals the crack resistance force.2 Griﬃth’s energy balance It is now assumed that the available external and internal energy is transferred into surface energy. Both surface energy Ua and internal energy Ui are plotted as a function of the crack length. This crack therefore cannot growth and is stable. the lefthand side of the equation is called the energy release rate G and the righthand side the crack resistance force R. a crack will grow. The released elastic energy is a quadratic function of the crack length a. The material volume. In that case the crack growth criterion is just met (G = R). the reaction force. An edge crack of length a is introduced in the plate. is indicated as the shaded area around the crack.
The crack is loaded in mode I by a nominal stress σ [Nm−2 ]. Therefore it is assumed that dUe = 0 during crack propagation. 4. According to Griﬃth’s energy balance. Because the edges with the applied stress are at a far distance from the crack. From this relation we can calculate the Griﬃth stress σgr and also the critical crack length ac . It is further assumed that the elastic energy in the elliptical area is released.3 Griﬃth stress We consider a crack of length 2a in an ”inﬁnite” plate with uniform thickness B [m]. dUa dUi < da da dUi dUa − > da da dUi dUa − = da da − → no crack growth → unstable crack growth → critical crack length 4. R 2γ ac available Ui ac Fig. .25 dUe = 0 a Ua da needed G. when the crack with length a is introduced.2 : Illustration of Griﬃth’s energy balance criterion. their displacement will be very small when the crack length changes slightly. The energy release rate and the crack resistance force can now be calculated. which is applied on edges at large distances from the crack. the applied stress σ and the crack length a are related.
26
σ
y thickness B 2a a x
σ
Fig. 4.3 : Elliptical region, which is unloaded due to the central crack of length 2a. Ui = 2πa2 B Griﬃth’s energy balance G=− 1 B dUi da =
2 1σ 2 E
; (dUe = 0)
Ua = 4aB γ
[Nm = J]
1 B
dUa da
=R
→
2πa
σ2 = 4γ E
[Jm−2 ]
Griﬃth stress
σgr =
2γE πa
critical crack length
ac =
2γE πσ 2
4.3.1
Discrepancy with experimental observations
When the Griﬃth stress is compared to the experimental critical stress for which a crack of length a will propagate, it appears that the Griﬃth stress is much too small: it underestimates the strength. The reason for this discrepancy is mainly due to the fact that in the Griﬃth energy balance the dissipation is neglected. This can be concluded from the comparison of the crack resistance force R = 4γ and the measured critical energy release rate Gc . For materials which are very brittle (e.g. glass) and thus show little dissipation during crack growth, the diﬀerence between R and Gc is not very large. For ductile materials, showing much dissipation (e.g. metal alloys), Gc can be 105 times R. Griﬃth’s energy balance can be used in practice, when the calculated energy release rate is compared to a measured critical value Gc . 1 B dUe dUi − da da
energy balance
G=
= R = Gc
27
critical crack length
ac =
Gc E 2πσ 2
Griﬃth’s crack criterion
G = Gc
4.4
Compliance change
The energy release rate can be calculated from the change in stiﬀness due to the elongation of a crack. It is common practice to use the compliance C instead of the stiﬀness and the relation between G and the change of C will be derived for a plate with an edge crack, which is loaded by a force F in point P . When the crack length increases from a to a + da, two extreme situations can be considered : • ﬁxed grips • constant load : point P is not allowed to move (u = 0), : force F is kept constant.
F u P a
F u P a + da
Fig. 4.4 : Edge crack in a plate loaded by a force F . F F a + da
a dUi
a
a + da dUi dUe
u u Fig. 4.5 : Internal and external energy for ﬁxed grips (left) and constant load (right).
28
4.4.1
Fixed grips
Using the ﬁxed grips approach, it is obvious that dUe = 0. The force F will decrease (dF < 0) upon crack growth and the change of the internally stored elastic energy can be expressed in the displacement u and the change dF . The energy release rate can be calculated according to its deﬁnition. ﬁxed grips : dUe = 0 dUi = Ui (a + da) − Ui (a) = = Griﬃth’s energy balance G=− 1 dF 1 u2 dC 1 2 dC u = = F 2 da 2B da 2B C 2B da
1 2 (F + 1 2 udF
(< 0)
dF )u −
1 2Fu
4.4.2
Constant load
With the constant load approach, the load will supply external work, when the crack propagates and the point P moves over a distance du. Also the elastic energy will diminish and this change can be expressed in F and du. The energy release rate can be calculated according to its deﬁnition. constant load dUe = Ue (a + da) − Ue (a) = F du dUi = Ui (a + da) − Ui (a) = = Griﬃth’s energy balance G= 1 du 1 2 dC F = F 2B da 2B da
1 2 F (u + 1 2 F du
(> 0)
du) −
1 2Fu
4.4.3
Experiment
It can be concluded that the energy release rate can be calculated from the change in compliance and that the result for the ﬁxed grip approach is exactly the same as that for the constant load method. In reality the loading of the plate may not be purely according to the extreme cases, as is shown in the ﬁgure. From such a real experiment the energy release rate can be determined from the forcedisplacement curve.
the crack length results in the energy release rate G. Using linear elastic beam theory. . the opening ∆u of the crack can be expressed in the force F . beam thickness B. 4.7 : Beam with a central crack loaded in Mode I.r.t. F u B 2h a u F Fig. the crack length a. which appears to be quadratic in a. G= shaded area 1 a4 − a3 B 4. 4. beam height 2h and Young’s modulus E. shown in the ﬁgure.6 : Experimental forcedisplacement curve during crack growth.4.4 Examples As an example the energy release rate is calculated for the socalled double cantilever beam. The compliance is the ratio of opening and force.29 F a1 a2 a3 a4 F u P a u Fig. Diﬀerentiating the compliance w.
so no function of the crack length. It is assumed that the height of the beam is an exponential function of a.8 : Cracked beam with variable crosssection. Applying again formulas from linear elastic beam theory.30 4F a3 F a3 = 3EI EBh3 2u 8a3 dC 24a2 ∆u = = → = C= F F EBh3 da EBh3 2 a2 1 1 2 dC 12F G= [J m−2 ] 2 F da = EB 2 h3 B u= Gc = 2γ → Fc = B a 1 3 6 γEh → The question arises for which beam geometry. h a Fig. 2u 8a3 ∆u = = F F EBh3 : h = h0 an → dC 24a2 = da EBh3 C= → choice u= C= dC da F a3 4F a3 4F a3(1−n) = = 3(1 − n)EI (1 − n)EBh3 (1 − n)EBh3 0 2u 8a3(1−n) = F (1 − n)EBh3 0 constant for n= → 2 3 dC 24a(2−3n) = da EBh3 0 h = h0 a 3 2 → . 4. the energy release rate will be constant. it can be derived for which shape G is no function of a.
but could also be determined. The ﬁrst sections present a summary of the theory of linear elasticity. This implies that relevant variables.r. 5.1 Deformation and strain The basic problem in mechanics is the prediction of the mechanical behavior of a body when it is subjected to an external load. For this purpose we use position vectors : X in the undeformed state and x in the deformed state. that it does not contain voids. 31 x + dx P x Q P . Throughout this chapter it is assumed that the material behavior is linear elastic and isotropic.e. i. The stress state is calculated using Airy stress functions. making the behavior thermomechanical. The displacement during deformation is obviously the diﬀerence between x and X. (Temperature is not considered here. are a continuous function of the location within the body. These components are indicated by indices.) To set up a mathematical model. In the mathematics of the next sections. In this way the diﬀerence vector between two adjacent points in undeformed and deformed state can be related. the deformation rate and the stresses in the material. 5. In continuum mechanics we assume that the body is perfect. like the displacement. which can be used to solve the above problem. Q X + dX u X e3 e2 e1 Fig.t. a coordinate system.1 : Deformation of a continuum. We want to predict the deformation. which take values in the rage {1. 2. 3} for threedimensional deformations. we use the components of the vectorial (and tensorial) variables w. cracks or other imperfections.Chapter 5 Stress concentrations In this chapter attention is given to stress concentrations due to holes. we must be able to identify the position of material points.
i uk.i + uk.13 = 0 2ε31.j + uj.22 = 0 2ε12.33 − ε23. these can be linearized to the linear strains εij .j + ui.22 − ε12.i ) Because three displacement components are the basis for the deﬁnition of six strain components.i + uk.k )dXk ] = (δij δik + δij ui. For small elongations and rotations.j )dXi dXj = dS 2 + (ui.k + uk.32 = 0 .j + uj.i + uk.23 + ε23.j )dXj The local deformation.21 = 0 ε33.j ui.j + uj.j )dXi dXj ds2 − dS 2 = (ui.j δik + ui.j ) 2 1 εij = 2 (ui.i uk.j ui.11 = 0 ε22.31 + ε31.i uk.i uk.j + uj.i + uk.j )dXi dXj = 2γij dXi dXj GreenLagrange strains linear strains γij = 1 (ui.33 − ε33.32 xi = Xi + ui (Xi ) xi + dxi = Xi + dXi + ui (Xi + dXi ) = Xi + dXi + ui (Xi ) + ui.i + uk.k )dXj dXk = (δij + ui. the GreenLagrange strains γij are commonly used.e.33 = 0 2ε23. ε11.11 − ε11. is described by strains.31 − ε33.23 − ε22. ds2 = dxi dxi = [(δij + ui.j dXj dxi = dXi + ui.12 − ε11. the deformation of a small material volume in a material point. When elongations and rotations are large. i.i uk. The compatibility relations express this dependency.12 + ε12.j + uj.31 − ε31.j dXj = (δij + ui.k + ui.12 − ε12.22 − ε22.j + uj. the latter cannot be independent.k )dXj dXk = (δjk + uj.j )dXi dXj = dXi dXi + (ui.23 − ε23.j )dXj ][(δik + ui.11 − ε31.
These stress components can be represented as the stress vectors on the sides of an inﬁnitesimal stress cube in the material point. Due to material symmetry – crystal structure – the elastic behavior of most materials can be described by considerably less material parameters. 2. j = 1. 5. For isotropic materials. two material parameters are suﬃcient : Young’s modulus E and Poisson’s ratio ν. can be calculated from the unity normal vector n on the plane and the Cauchy stress tensor. The relation is given by the stiﬀness parameters Cijkl (i. 3). The stress vector p on a plane in a material point. k. with its sides normal to the coordinate axes. the components of p can be related to those of n by the Cauchy stress components σij (i. unity normal vector stress vector Cauchy stress components n = ni ei p = pi ei pi = σij nj σ33 σ23 σ13 3 2 1 σ11 σ31 σ21 σ12 σ32 σ22 Fig. σij = Cijkl εlk .33 5.2 : Stress components on a stress cube. 5. it provokes stresses in the material. Using index notation in a coordinate system.2 Stress When deformation is not completely unconstrained. Due to symmetry of stress and strain components. j. 3) and is referred to as Hooke’s law. the mechanical behavior of a solid material is reversible and characterized by a linear relation between stress and strain components. there are no more than 21 independent stiﬀness parameters. l = 1. 2.3 Linear elastic material behavior For very small deformations.
which is subjected to external loads. Hooke’s law is represented by a sixbysix matrix.34 Hooke’s law for isotropic materials For isotropic materials. 3 i = 1. relating six strain components to six stress components and vice versa. will deform and the stresses in the deformed state have to satisfy the equilibrium equations in each point. 3 σ11 σ22 σ33 σ12 σ23 σ31 1−ν ν ν 0 0 0 ν 1−ν ν 0 0 0 ν ν 1−ν 0 0 0 = α 0 0 0 1 − 2ν 0 0 0 0 0 0 1 − 2ν 0 0 0 0 0 0 1 − 2ν α = E/[(1 + ν)(1 − 2ν)] 1 −ν −ν 0 0 0 −ν 1 −ν 0 0 0 −ν −ν 1 0 0 0 0 0 0 1+ν 0 0 0 0 0 0 1+ν 0 0 0 0 0 0 1+ν σ11 σ22 σ33 σ12 σ23 σ31 ε11 ε22 ε33 ε12 ε23 ε31 ε11 ε22 ε33 ε12 ε23 ε31 = 1 E 5. 2. . E ν δij εkk εij + 1+ν 1 − 2ν ν 1+ν δij σkk σij − εij = E 1+ν σij = i = 1. When these components are stored in columns. We have assumed here that the acceleration of the material points can be neglected. Hooke’s law can be written in index notation. The three equilibrium equations for forces are partial diﬀerential equations. 2.4 Equilibrium equations A material. which can only be solved with proper boundary conditions. The three equilibrium equations for moments indicate that the Cauchy stress components are symmetric.
volume load force equilibrium moment equilibrium ρqi σij. They have to satisfy two equilibrium equations.1 dx1 3 σ11 σ21 σ31 σ12 + σ12. 3 5.3 dx3 σ33 + σ33. equilibrium (qi = 0) compatibility Hooke’s law σij = E 1+ν εij + ν δij εkk 1−ν . which we take to be the {1. 5. The three strain components in the plane are related by one compatibility relation.j + ρqi = 0 σij = σji i = 1.12 − ε11. 2.3 : Stress components on a stress cube with a volume load.35 σ13 + σ13. σ21.1 + σ12. 2 σ11.2 = 0 . with a 3 × 3 stiﬀness matrix or compliance matrix.3 dx3 σ23 + σ23. we have a plane stress state.2 dx2 σ32 + σ32. Only the three independent stress components in the plane have to be considered.1 + σ22. 2}plane in this case.22 − ε22.1 dx1 σ21 + σ21.2 = 0 2ε12.11 = 0 Hooke’s law in matrix notation .2 dx2 2 σ13 σ23 σ33 1 Fig.3 dx3 σ12 σ22 σ32 σ11 + σ11.1 dx1 σ31 + σ31. εij = 1+ν E σij − ν δij σkk 1+ν i = 1. Hooke’s law for linear elastic material behavior relates the three stress components to the three strain components.2 dx2 σ22 + σ22.5 Plane stress When stresses are zero on a certain plane.
2 σ11. Only three strain components remain. perpendicular to the {1.36 1 ε11 ε22 = 1 −ν E 0 ε12 σ11 σ22 = E 1 − ν2 σ12 ε33 = − ε13 −ν 0 σ11 1 0 σ22 0 1+ν σ12 1 ν 0 ε11 ν 1 0 ε22 0 0 1−ν ε12 ν ν (σ11 + σ22 ) = − (ε11 + ε22 ) E 1−ν = ε23 = 0 5.1 + σ22.12 − ε11. which have to satisfy one compatibility relation. σij = E 1+ν εij + ν δij εkk 1 − 2ν i = 1.2 = 0 .6 Plane strain In the plane strain situation. 2}plane.1 + σ12. Hooke’s law can be simpliﬁed and represented with 3 × 3 stiﬀness and compliance matrices. equilibrium (qi = 0) compatibility Hooke’s law εij = 1+ν (σij − νδij σkk ) E .22 − ε22. Three stress components have to be solved from two equilibrium equations.11 = 0 Hooke’s law in matrix notation σ11 1−ν ν 0 ε11 E σ22 = ν ε22 1−ν 0 (1 + ν)(1 − 2ν) σ12 0 0 1 − 2ν ε12 1 − ν −ν 0 σ11 ε11 ε22 = 1 + ν −ν 1 − ν 0 σ22 E 0 0 1 σ12 ε12 σ33 = σ13 Eν (ε11 + ε22 ) = ν (σ11 + σ22 ) (1 + ν)(1 − 2ν) = σ23 = 0 . σ21. the deformation of the material is such that there is no elongation and shear in and perpendicular to one direction.2 = 0 2ε12. which we here take to be the 3direction.
8 Stress function method According to the displacement method. The result is a fourthorder diﬀerential equation for the stress function ψ. so only one compatibility equation is relevant.j + 1 − 2ν (ui. With proper boundary conditions (4). the strain components are expressed in derivatives of ψ.ij + δij ψ. The stress components are related to derivatives of the stress function.j 1 − 2ν =0 ν δij εkk. ψ(x1 . the equilibrium equation is transformed to a diﬀerential equation in the displacements. the equilibrium is transformed to a diﬀerential equation in the Airy stress function ψ. which is called the biharmonic equation. It can be written in short form using the Laplace operator ∇2 . Analytical solutions can only be determined for simple cases.ij ) + δij uk.kk → σij.j = 0 1+ν (σij − νδij σkk ) E .37 5. the stressstrain relation according to Hooke’s law are substituted into the equilibrium equations. the biharmonic equation can be solved (theoretically). When the stress function is known.j εij + ν δij εkk. the stresses. Next. Here only twodimensional problems are considered.7 Displacement method The displacement method is based on transforming the equilibrium equations to diﬀerential equations for the displacement components. In the socalled stress function method. First. strains and displacements can be derived rather straightforwardly. resulting in the ﬁnal two diﬀerential equations for the two displacement components. the straindisplacement relations are substituted.i ) (ui.j = 0 σij = E 1+ν εij = E 1+ν BC’s ui → εij → σij 1 2 1 2 E 1+ν εij. x2 ) εij = → σij = −ψ. leading to two diﬀerential equations for the strain components. Using Hooke’s law.jj Eν + uj.j + uj.kj = 0 (1 + ν)(1 − 2ν) 5. These equations can be solved with proper boundary conditions (BC’s). The strain components are then substituted in the compatibility equations. such that they satisfy the equilibrium equations. σij.
12 − ε11.1122 + ψ.2222 + ψ.22 = 0 Laplace operator : ∂2 ∂2 + 2 = ( )11 + ( )22 ∇ = ∂x2 ∂x2 1 2 → biharmonic equation BC’s ψ → σij → εij → ∇2 (∇2 ψ) = ∇4 ψ = 0 ui Cylindrical coordinates The biharmonic equation has been derived and expressed in Cartesian coordinates.38 εij = 1+ν {−ψ.22 ).11 = 0 2ψ.22 ). e3 } → {er .11 + ψ. et .22 − ε22. It is immediately clear that cylindrical base vectors are a function of the cylindrical coordinate θ.1111 = 0 → (ψ. 5. e2 .11 + (ψ. To get an expression in cylindrical coordinates.ij + (1 − ν)δij ψ. the base vectors of the cylindrical coordinate system are expressed in the base vectors of the Cartesian coordinate system.kk } E 2ε12.4 : Cartesian and cylindrical coordinate system. vector bases {e1 . ez } er = er (θ) = e1 cos θ + e2 sin θ et = et (θ) = −e1 sin θ + e2 cos θ .11 + ψ. z ez et er e3 e2 e1 θ x r y Fig.
gradient operator Laplace operator twodimensional ∂ 1 ∂ ∂ + et + ez ∂r r ∂θ ∂z 2 1 ∂2 1 ∂ ∂2 ∂ + 2 + 2 ∇2 = ∇ · ∇ = 2 + 2 ∂r r ∂r r ∂θ ∂z 2 2 1 ∂ 1 ∂ ∂ + 2 ∇2 = 2 + ∂r r ∂r r ∂θ 2 ∇ = er Biharmonic equation The biharmonic equation can now be written in cylindrical coordinates. the biharmonic equation must be solved to determine the Airy stress function ψ(r. To determine the stress components as a function of r and θ. The latter expression can best be derived from the expression for the stress tensor and using the expression for the gradient operator in cylindrical coordinates. . ∂ {et (θ)} = −er (θ) ∂θ Laplace operator The Laplace operator is a scalar operator and can be written as the inner product of the gradient operator. θ) for this problem. which are derivatives of the stress function. which is a vector operator.39 ∂ {er (θ)} = et (θ) ∂θ . The center of the hole is the origin of a cylindrical coordinate system. The plate is loaded with a uniform stress σ in the global xdirection. biharmonic equation ∂2 1 ∂2 1 ∂ + 2 2 + ∂r 2 r ∂r r ∂θ stress components σrr = 1 ∂ψ 1 ∂2ψ + 2 2 r ∂r r ∂θ ∂2ψ σtt = ∂r 2 ∂ 1 ∂ψ 1 ∂ 2 ψ − =− σrt = 2 r ∂θ r ∂r∂θ ∂r ∂ 2 ψ 1 ∂ψ 1 ∂2ψ + 2 2 + ∂r 2 r ∂r r ∂θ =0 1 ∂ψ r ∂θ 5. Dimensions of the plate are much larger than 2a. Also the stress components.9 Circular hole in ’inﬁnite’ plate A circular hole with diameter 2a is located in a plate with uniform thickness. can be written in cylindrical coordinates.
5.40 y σ σ r θ x 2a Fig. loaded with a uniform stress σ in xdirection.5 : A circular hole in a plate. θ) = − 1 σ sin(2θ) 2 . which can be related to σ using equilibrium. the uniform stress σ can be decomposed in a radial stress σrr and a shear stress σrt . 1 σrr (r = b. 5. σ σ θ 2a σrr σrt σrr σrt 2b Fig. Load transformation At a circular crosssection at a distance r = b ≫ a.6 : Transformation of the xload in radial and circumferential loads. The total load at the crosssection r = b is the sum of a radial load and a load that depends on θ. θ) = 2 σ + 1 σ cos(2θ) 2 σrt (r = b. the stress ﬁeld for b ≤ r ≤ a will be determined. For both loadcases. after which the combined solution follows from superposition.
the problem is axisymmetric and ψ is only a function of r. Airy function stress components 1 ∂ψ 1 ∂2ψ 1 df + 2 2 = r ∂r r ∂θ r dr d2 f ∂2ψ = 2 σtt = ∂r 2 dr ∂ 1 ∂ψ σrt = − =0 ∂r r ∂θ σrr = biharmonic equation 1 d d2 + 2 dr r dr d2 f 1 df + 2 dr r dr =0 ψ = f (r) A general solution for the diﬀerential equation has four integration constants. which relates the strain components εrr and εtt . σrt (r = b) = − 1 σ sin(2θ) 2 σrr (r = a) = σrt (r = a) = 0 1 σrr (r = b) = 2 σ cos(2θ) Load case I In the ﬁrst loadcase. However. σrt (r = b) = 0 . the boundary at r = b is subjected to a radial load σrr = 1 σ. The biharmonic equation is an ordinary diﬀerential equation in the variable r. σrr (r = a) = σrt (r = a) = 0 1 σrr (r = b) = 2 σ . general solution stresses ψ(r) = A ln r + Br 2 ln r + Cr 2 + D A + B(1 + 2 ln r) + 2C r2 A σtt = − 2 + B(3 + 2 ln r) + 2C r σrr = strains (from Hooke’s law for plane stress) εrr = 1 A (1 + ν) + B{(1 − 3ν) + 2(1 − ν) ln r} + 2C(1 − ν) E r2 11 A εtt = − (1 + ν) + B{(3 − ν)r + 2(1 − ν)r ln r} + 2C(1 − ν)r Er r . Because 2 σrr is not depending on θ. the boundary conditions for σrt cannot be used as this stress component is zero everywhere. II. The stress components can be expressed in this function. Also b ≫ a is used in the calculation.41 two load cases I. Instead the compatibility equation is used. These have to be determined from the boundary conditions for this loadcase.
r ∂r r ∂θ 1 ∂ψ 1 ∂ 2 ψ ∂ σrt = 2 − =− r ∂θ r ∂r∂θ ∂r σrr = biharmonic equation d2 g 1 dg 1 ∂2 4 1 ∂ ∂2 + 2 2 − g cos(2θ) + + 2 ∂r r ∂r r ∂θ dr 2 r dr r 2 d2 d2 g 1 dg 4 4 1 d − 2 − 2 g cos(2θ) = 0 + + 2 2 dr r dr r dr r dr r =0 → σtt = 1 ∂ψ r ∂θ ∂2ψ ∂r 2 ψ(r.) Substitution in the biharmonic equation results in a diﬀerential equation for g(r) and expressions for the stress components. θ) must be written as the product of a function g(r) and cos(2θ). θ.r.42 compatibility 2 BC’s and b ≫ a εrr = → du d(r εtt ) = dr dr a2 ) r2 → . Confronting the general expressions for stresses σrr and σrt with these prescribed values at r = b. general solution g = Ar 2 + Br 4 + C 1 +D r2 → ψ = Ar 2 + Br 4 + C r1 + D cos(2θ) 2 6C 4D + 2 4 r r 6C 2A + 12Br 2 + 4 r stresses σrr = − 2A + σtt = cos(2θ) cos(2θ) . These have to be determined from the boundary conditions for this loadcase. At the boundary r = b the stresses are harmonic functions of the coordinate θ. (The term cos(2θ) remains after two subsequent diﬀerentiations w. σrt = 0 Load case II The second loadcase is again deﬁned by the stress components at r = a and r = b. Airy function stress components 1 ∂2ψ 1 ∂ψ + 2 2 . the stress components σrr . → a2 ) r2 B=0 A and C 1 σrr = 2 σ(1 − 1 σtt = 2 σ(1 + . σtt and σrt can be determined as a function of r and θ. indicates that the Airy function ψ(r. θ) = g(r) cos(2θ) A general solution for the diﬀerential equation has four integration constants.t. Using b ≫ a.
θ) = σrt (r. The stress concentration factor (SCF) Kt is a ratio of the maximum stress and the applied stress and therefore it is a dimensionless number. This makes a larger hole to be more dangerous regarding the occurrence of damage than the smaller hole. σrr (r = a.43 σrt = 2A + 6Br 2 − 2D 6C − 2 r4 r → − sin(2θ) 4 BC’s and b ≫ a → A. . θ) = σrt (r = a. it appears that the stress gradient is higher when the hole diameter is smaller. σ 2 σ σtt = 2 3a4 4a2 a2 + 1 + 4 − 2 cos(2θ) r2 r r 4 2 3a a 1 + 2 − 1 + 4 cos(2θ) r r 2 4 σ 2a 3a =− 1 − 4 + 2 sin(2θ) 2 r r 1− σrr = σrt Special points At the inner hole boundary (r = a). It appears not to depend on the hole diameter. θ = 0) = 0 σtt (r = a. σtt and σrt resulting from the total load at r = b ≫ a. the stress σtt is maximum for θ = π and the stress 2 concentration is 3.B.C and D σrr = 1 σ 1 + 2 cos(2θ) r4 3a4 σtt = − 1 σ 1 + 4 cos(2θ) 2 r 3a4 2a2 σrt = − 1 σ 1 − 4 + 2 sin(2θ) 2 r r 3a4 4a2 r2 Stresses for total load The stress components σrr . which means that it is three times the applied stress σ. The chance that there is a defect or ﬂaw in this volume is higher than in the specimen with the smaller hole. For a larger hole there is more volume of material with a higher than nominal stress level. θ = π ) = 3σ 2 σtt (r = a. θ = 0) = −σ stress concentration factor Kt = σmax =3 σ [] Stress gradients When we plot the stress σtt as a function of the distance to the hole. are the sum of the solutions for the separate loadcases.
the stress concentration becomes inﬁnite when the radius approaches zero.7 : Diﬀerent stress gradient at holes with diﬀerent radius. The solution reveals that at the location of the smallest radius ρ.8 : Elliptical hole loaded with a uniform stress σ in ydirection. a = σ 1 + 2 a/ρ ≈ 2σ b [] σyy (x = a. as would be the situation for a real crack. 5. y σyy σ a b σ x radius ρ Fig.44 Fig. y = 0) = σ 1 + 2 stress concentration factor a/ρ Kt = 2 a/ρ . 5.10 Elliptical hole Stress components in the vicinity of an elliptical hole can be calculated also. 5.
It is obvious that we can 2 write : ei = ier . a real crack has zero radius (ρ = 0) and such a crack tip represents a singularity. 6. stresses. 6. Due to this singularity.1 : Line crack with local coordinate systems originating in the crack tip. II or III. However. 45 . which are located on the real and imaginary axis of the complex coordinate system. complex variables and functions will enter the mathematics.1 Complex plane The crack tip is considered to be the origin of a complex coordinate system with real axis x1 and imaginary axis x2 . It is easily seen that multiplying z with the imaginary constant i results in a counterclockwise rotation over π of vector z. The position in this complex plane can be identiﬁed by a complex number and all variables – Airy function.1. strains.Chapter 6 Crack tip stresses When a linear crack is loaded in ModeI. x2 r θ x1 Fig. The conjugate complex number z ¯ is mirrored with respect to the real axis. stress components at the crack tip can be calculated with the Airy stress function method. allows the complex numbers to be represented as a vector z. 6. ¯ Introducing unity vectors er and ei along the real and imaginary axis. displacements – are considered to be complex functions of this position. The real and imaginary part of z can be expressed in z and z .1 Complex variables A complex number z has a real (x1 ) and an imaginary (x2 ) part.
2 Complex functions A scalar function of a complex number z results in a complex number f .1. x2 ) − iζ(x1 . In Appendix B a detailed ¯ derivation is presented. f (z) = φ + iζ = φ(x1 . It can be shown (see Appendix A) that both the real and number f ¯ the imaginary part of f (and f ) satisfy the Laplace equation.1.46 x2 ei er r θ x1 z z ¯ Fig. x2 = . resulting in a Laplace operator in terms of these variables. These derivatives can be transformed to derivatives with respect to the variables z and z . We only consider functions which can be written as the sum of a real (φ) and an imaginary (ζ) part. z = x1 − ix2 = re−iθ ¯ 1 2i (z z = x1 er + x2 ei = x1 er + x2 ier = (x1 + ix2 )er − z ) = − 1 i(z − z ) ¯ ¯ 2 6. In Cartesian coordinates this operator comprises secondorder derivatives with respect to the Cartesian coordinates x1 and x2 . x2 ) = f z ¯ φ = 1 {f + f } 2 . .2 : Complex plane and two complex numbers. the result is a complex ¯ ¯. which is conjugate to f . x2 ) = f ¯ f (¯) = φ(x1 . x2 ) + iζ(x1 . 1 ¯ ζ = − 2 i{f − f } → 6. 6. When the conjugate complex number z is substituted in the scalar function. z = x1 + ix2 = reiθ 1 x1 = 2 (z + z ) ¯ .3 Laplace operator The biharmonic equation includes the Laplace operator ∇2 .
x2 ) = g(z.47 complex function Laplacian derivatives g(x1 .4 Biharmonic equation The Airy stress function ψ is taken to be a function of z and z and it must be solved from ¯ the biharmonic equation. z ) = ∇2 (φ(z. Thus.1.2 Solution of biharmonic equation According to subsection 6. from the biharmonic equation it follows that the Airy function must satisfy a Poisson equation with righthand side the sum of a complex function and its conjugate. z )) = 0 ¯ ¯ ∇2 ψ = 4 ∂2ψ ¯ =φ=f +f → ∂z∂ z ¯ → ψ= ¯ φ=f +f → 1 2 ¯ zΩ + zΩ + ω + ω ¯ ¯ . z ) ¯ ∇2 g = ∂2g ∂2g + 2 ∂x2 ∂x2 1 ∂g ¯ ∂g ∂z ∂g ∂ z ∂g ∂g = + = + ∂x1 ∂z ∂x1 ∂ z ∂x1 ¯ ∂z ∂ z ¯ 2g 2g 2g 2g ∂ ∂ ∂ ∂ 2 = ∂z 2 + 2 ∂z∂ z + ∂ z 2 ¯ ¯ ∂x1 ∂g ¯ ∂g ∂g ∂z ∂g ∂ z ∂g −i = + =i ∂x2 ∂z ∂x2 ∂ z ∂x2 ¯ ∂z ∂z ¯ 2g 2g 2g 2g ∂ ∂ ∂ ∂ 2 = − ∂z 2 + 2 ∂z∂ z − ∂ z 2 ¯ ¯ ∂x2 ∂2g ∂2g ∂2g + 2 =4 ∂z∂ z ¯ ∂x2 ∂x2 1 ∂2 ∂z∂ z ¯ Laplacian ∇2 g = → ∇2 = 4 6. z ) = 0 ¯ 6. The general solution for ψ contains two unknown ¯ complex functions Ω and ω and there conjugates Ω and ω. the solution of a Laplace equation can be considered to be the real part of a complex function.2 (and Appendix A). z ) ¯ ∇2 ∇2 ψ(z. Airy function biharmonic equation ψ(z. ¯ ∇2 ∇2 ψ(z.1.
ω. The detailed ¯ ¯ derivation can be found in Appendix B.11 + ψ. z ) ¯ u = u1 − iu2 = u(z.48 6. The real part is the x1 component and the imaginary part is the x2 component. ∂z ¯ ∂z .11 = Ω ′ + Ω ′ + → 1 2 1 2 ¯ z Ω ′′ + ω ′′ + z Ω ′′ + ω ′′ ¯ ¯ ′′ ′′ ¯ ′′ + ω ′′ zΩ + ω + zΩ ¯ ¯ 1 ¯ σ12 = −ψ. z and z are indicated with ( )′ . Considering the coordinate axes to be real and imaginary axes in a complex plane.t. x2 ) + iu2 (x1 .r. we have e1 = er and e2 = ei = ier .γγ = ψ.2. 6.r. x2 } with orthonormal base vectors e1 and e2 .r.γγ = ψ.3 : Complex displacement. z and z ¯ ¯ ¯ ¯ can be determined. which is a complex number. z ) = −ψ.t. a Cartesian coordinate system {x1 .1 Stresses From the general solution for ψ the stress components σ11 . z ) ¯ ¯ ¯ → Because the displacements u and u are a function of z and z . x2 ) = u(z. the derivatives w.2 Displacement The displacement vector u can be written in components w.kk ¯ ¯ σ11 = −ψ.22 = Ω ′ + Ω ′ − ¯ σ22 = −ψ. σ22 and σ12 can be derived and ¯ expressed in Ω.t.2. The displacement vector can thus be considered to be a vector in the direction er .22 + ψ. σij = σij (z. u u2 x2 ei e2 er e1 r θ x1 u1 Fig. Ω and ω.12 = − 2 i z Ω ′′ + ω ′′ − z Ω ′′ − ω ′′ ¯ ¯ 6. Derivatives w. It appears that ∂u and ∂ u can be expressed in strain components only.ij + δij ψ. u = u1 e1 + u2 e2 = u1 er + u2 ei = u1 er + u2 ier = (u1 + iu2 )er = uer u = u1 + iu2 = u1 (x1 . having a length u.
the derivatives of these can be determined ¯ ¯ and added. Integration leads to an expression ¯ for u.49 ∂u ∂u ∂x1 ∂u ∂x2 1 = + =2 ∂z ¯ ∂x1 ∂ z ¯ ∂x2 ∂ z ¯ ∂u2 ∂u1 ∂u1 1 +i +i − = 2 ∂x1 ∂x1 ∂x2 ∂u ∂x1 ∂u ∂x2 ∂u 1 = + =2 ∂z ∂x1 ∂z ∂x2 ∂z ∂u1 ∂u2 ∂u1 1 = 2 +i −i + ∂x1 ∂x1 ∂x2 ∂u ¯ ∂ u ∂x1 ¯ ∂ u ∂x2 ¯ 1 = + =2 ∂z ∂x1 ∂z ∂x2 ∂z ∂u2 ∂u1 ∂u1 1 −i −i − = 2 ∂x1 ∂x1 ∂x2 ∂ u ∂x1 ¯ ∂ u ∂x2 ¯ ∂u ¯ 1 = + =2 ∂z ¯ ∂x1 ∂ z ¯ ∂x2 ∂ z ¯ ∂u1 ∂u2 ∂u1 1 = 2 −i +i + ∂x1 ∂x1 ∂x2 ∂u ∂x1 ∂u2 ∂x2 ∂u ∂x1 ∂u2 ∂x2 ∂u ¯ ∂x1 ∂u2 ∂x2 ∂u ¯ ∂x1 ∂u2 ∂x2 +i ∂u ∂x2 1 2 = −i (ε11 − ε22 + 2iε12 ) ∂u ∂x2 1 2 = −i ε11 + ε22 + i ∂u2 ∂u1 − ∂x1 ∂x2 ∂u ¯ ∂x2 1 2 = +i (ε11 − ε22 − 2iε12 ) ∂u ¯ ∂x2 1 2 = ε11 + ε22 − i ∂u2 ∂u1 − ∂x1 ∂x2 Using Hooke’s law for plane strain. 1+ν E 1+ν =− E 1+ν =− E 1+ν =− E u=− ∂u ∂z u ¯ ∂u ¯ ∂z ¯ ¯ zΩ ′ + ω′ + M ¯ ¯ Ω′ + M ′ ¯ zΩ ′ + ω′ + M ¯ ¯ Ω′ + M ′ . ∂u = ∂z ¯ 1+ν σ11 − σ22 + 2iσ12 E 1+ν ¯ =− z Ω ′′ + ω ′′ ¯ E 1 2 u=− 1+ν ¯ zΩ ′ + ω′ + M ¯ E Using the general solution for u and the conjugate u. comprising an unknown function M (z).r. these derivatives can also be calculated. z can be expressed in the stress ¯ ¯ components and thus in the complex functions Ω and ω . the derivative of u w. From the general solution. Adding the directly derived derivatives ∂u and ∂ u leads again to an expression ∂z ∂z ¯ ¯ in strain components and through Hooke’s law in the functions Ω ′ and Ω ′ .t. whereupon the integration function M can be determined.
respectively. ¯ The displacement appears to be proportional to r λ+1 . the displacement u(z.3 Choice of complex functions Two unknown functions Ω and ω have to be determined from boundary conditions. are functions of the variables r and θ and the constants α. This argument requires λ to be larger than −1. Euler’s formula is used to express u in cosine and sine functions. δ and λ. still containing ﬁve unknown constants : α. Using these functions and their derivatives.2. however. From a physical point of view.4 Displacement components The displacement u is a complex function. γ. δ and λ. γ. .2. Its real and imaginary part. the displacement always has to be ﬁnite. Ω = (α + iβ)z λ+1 = (α + iβ)r λ+1 eiθ(λ+1) ω ′ = (γ + iδ)z λ+1 = (γ + iδ)r λ+1 eiθ(λ+1) → ¯ Ω = (α − iβ)¯λ+1 = (α − iβ)r λ+1 e−iθ(λ+1) z ¯ ′ = (α − iβ)(λ + 1)¯λ = (α − iβ)(λ + 1)r λ e−iθλ Ω z ω ′ = (γ − iδ)¯λ+1 = (γ − iδ)r λ+1 e−iθ(λ+1) ¯ z 1 λ+1 κ(α + iβ)eiθ(λ+1) − r u= 2µ E 2(1 + ν) → (α − iβ)(λ + 1)eiθ(1−λ) − (γ − iδ)e−iθ(λ+1) with µ= displacement ﬁnite → λ > −1 6. β. z ) can be derived. These two functions are speciﬁed to be of a certain form. β. u1 and u2 .50 ∂u ∂ u ¯ 1 + ν ¯′ ¯ + =− Ω + Ω′ + M ′ + M ′ ∂z ∂z ¯ E ∂u ∂ u ¯ 1+ν + = ε11 + ε22 = [(1 − 2ν)(σ11 + σ22 )] ∂z ∂z ¯ E (1 + ν)(1 − 2ν) ¯ = 2 Ω′ + Ω′ E ¯ ¯ M ′ + M ′ = −(3 − 4ν) Ω ′ + Ω ′ → M = −(3 − 4ν)Ω = −κΩ u=− 1+ν ¯ z Ω ′ + ω ′ − κΩ ¯ E 6.
3. → β=δ=0 −→ ω ′ = γz λ+1 = γr λ+1 ei(λ+1)θ 6. This can be concluded when studying the displacement components and using sin(−θ) = − sin(θ) and cos(−θ) = cos(θ).t. which is the crack direction. γ and λ. allow the determination of the integration constants. 6. To meet this requirement. u1 (θ > 0) = u1 (θ < 0) u2 (θ > 0) = −u2 (θ < 0) Ω = αz λ+1 = αr λ+1 ei(λ+1)θ .3. the x1 axis. In the functions Ω and ω ′ three unknown constants remain.r. the constants β and δ have to be zero.2 Stress components The stress components σ11 .3 Mode I For a crack loaded in Mode I.51 1 λ+1 u= κ(α + iβ)eiθ(λ+1) − r 2µ (α − iβ)(λ + 1)eiθ(1−λ) − (γ − iδ)e−iθ(λ+1) eiθ = cos(θ) + i sin(θ) u= 1 λ+1 r 2µ κα cos(θ(λ + 1)) − κβ sin(θ(λ + 1)) − α(λ + 1) cos(θ(1 − λ)) − β(λ + 1) sin(θ(1 − λ)) − γ cos(θ(λ + 1)) + δ sin(θ(λ + 1)) i + κα sin(θ(λ + 1)) + κβ cos(θ(λ + 1)) − α(λ + 1) sin(θ(1 − λ)) + β(λ + 1) cos(θ(1 − λ)) + γ sin(θ(λ + 1)) + δ cos(θ(λ + 1)) = u1 + iu2 6. symmetry of the normal displacement and boundary conditions for crack face stresses. .1 Displacement For Mode I loading of the crack. σ22 and σ12 can be expressed in the variables r and θ and the constants α. the displacement must be symmetric w.
The stress ﬁeld is a series of terms. 2. eiθ − e−iθ = 2i sin(θ) → σ11 = 2(λ + 1)r λ α cos(λθ) + 1 {αλ cos((λ − 2)θ) + γ cos(λθ)} 2 σ22 = 2(λ + 1)r λ α cos(λθ) − 1 {αλ cos((λ − 2)θ) + γ cos(λθ)} 2 σ12 = (λ + 1)r λ [αλ sin((λ − 2)θ) + γ sin(λθ)] The boundary condition stating that the crack surface is stressfree. The 2 2 ﬁrst terms are presented here.. For each value of λ we ﬁnd a ratio between α and γ. 1 . 1. 0. · · ·]. one for each value of λ in the range [− 1 . results in two equations for the unknown constants α and γ. . This conﬁnes the values of λ to be a series of discrete numbers.52 σ11 = (λ + 1) αz λ + α¯λ − z σ22 = (λ + 1) αz λ + α¯λ + z 1 2 1 2 αλ¯z λ−1 + γz λ + αλz¯λ−1 + γ z λ z z ¯ αλ¯z λ−1 + γz λ + αλz¯λ−1 + γ z λ z z ¯ σ12 = − 1 i(λ + 1) αλ¯z λ−1 + γz λ − αλz¯λ−1 − γ z λ z z ¯ 2 with z = reiθ . Only one unknown constant (γ) remains in these solution. 1 λ = −2 1 λ= 2 → → α = 2γ α = −2γ 3 1 σ11 = 2γr − 2 cos( 1 θ) 1 − sin( 2 θ) sin( 2 θ) + · · · 2 1 1 1 . . λ=0 λ=1 → → α = 1γ 2 α=γ 3 1 σ22 = 2γr − 2 cos( 1 θ) 1 + sin( 2 θ) sin( 2 θ) + · · · 2 1 3 σ12 = 2γr − 2 cos( 2 θ) cos( 2 θ) sin( 1 θ) + · · · 2 . σ22 (θ = ±π) = σ12 (θ = ±π) = 0 (λ − 2) cos(λπ) cos(λπ) λ sin(λπ) sin(λπ) det (λ − 2) cos(λπ) cos(λπ) λ sin(λπ) sin(λπ) with α γ −→ = 0 0 → = − sin(2λπ) = 0 → 2πλ = nπ → 1 n λ = −2. 2 n = 0. . z = re−iθ ¯ σ11 = (λ + 1)r λ αeiλθ + αe−iλθ − 1 2 → αλei(λ−2)θ + γeiλθ + αλe−i(λ−2)θ + γe−iλθ σ22 = (λ + 1)r λ αeiλθ + αe−iλθ + 1 2 αλei(λ−2)θ + γeiλθ + αλe−i(λ−2)θ + γe−iλθ 1 σ12 = − 2 i(λ + 1)r λ αλei(λ−2)θ + γeiλθ − αλe−i(λ−2)θ − γe−iλθ with eiθ + e−iθ = 2 cos(θ) . For a nontrivial solution the equations need to be dependent. 1.
SIF or KI . It can be calculated when more boundary conditions are available in a situation where geometry and load are speciﬁed in more detail.53 6. This term represents the singularity of the stress ﬁeld. A new constant is introduced in the stress ﬁeld. This can be concluded from the displacement components. The SIF KI is the only unknown constant in this solution.4. allow the determination of the integration constants. when approaching the crack tip. . √ √ 2πr σ22 θ=0 = 2γ 2π 1 KI = lim r →0 [ m 2 N m−2 ] 6. because it will become inﬁnite. The displacement components are also known and for λ = − 1 they are proportional to 2 √ r. the socalled Stress Intensity Factor.3. It is deﬁned as the limit value of the 22component of the crack tip stress.4 Mode II For a crack loaded in Mode II. using sin(−θ) = − sin(θ) and cos(−θ) = cos(θ).3. In the functions Ω and ω ′ three unknown constants remain. 6. antisymmetry of the tangential displacement component and boundary conditions for the crack face stresses.1 Displacement For Mode II loading of the crack. which is the crack direction. To meet this requirement.r. KI 1 cos( 2 θ) 1 − sin( 1 θ) sin( 3 θ) σ11 = √ 2 2 2πr KI 1 σ22 = √ cos( 2 θ) 1 + sin( 1 θ) sin( 3 θ) 2 2 2πr KI 1 1 cos( 2 θ) sin( 2 θ) cos( 3 θ) σ12 = √ 2 2πr KI 2µ KI u2 = 2µ u1 = r cos( 1 θ) κ − 1 + 2 sin2 ( 1 θ) 2 2 2π r sin( 1 θ) κ + 1 − 2 cos2 ( 1 θ) 2 2 2π 6.3 Stress intensity factor The ﬁrst term in the series is the dominant term close to the crack tip.t. the constants α and γ have to be zero. the displacement must be antisymmetric w. For plane stress and plane strain they diﬀer because of the diﬀerent versions of Hooke’s law. the x1 axis.4 Crack tip solution The stress components are determined for Mode I loading of a straight crack.
For plane stress and plane strain they diﬀer because of the diﬀerent versions of Hooke’s law.54 u1 (θ > 0) = −u1 (θ < 0) u2 (θ > 0) = u2 (θ < 0) Ω = iβz λ+1 = iβr λ+1 ei(λ+1)θ → .4.2 Stress intensity factor The stress intensity factor KII is now deﬁned as the limit value of the 12component of the crack tip stress. α=γ=0 −→ ω ′ = iδz λ+1 = iδr λ+1 ei(λ+1)θ 6.4.3 Crack tip solution The stress components are determined for Mode II loading of a straight crack. only the outofplane displacement u3 is relevant . The SIF KII is the only unknown constant in this solution. The displacement components are also known and for λ = − 1 they are proportional to 2 √ r. KII 3 1 σ11 = √ − sin( 2 θ) 2 + cos( 1 θ) cos( 2 θ) 2 2πr KII 3 1 σ22 = √ sin( 2 θ) cos( 1 θ) cos( 2 θ) 2 2πr KII 3 1 cos( 2 θ) 1 − sin( 1 θ) sin( 2 θ) σ12 = √ 2 2πr KII 2µ KII u2 = 2µ u1 = r sin( 1 θ) κ + 1 + 2 cos2 ( 1 θ) 2 2 2π r 1 − cos( 2 θ) κ − 1 − 2 sin2 ( 1 θ) 2 2π 6. KII = lim r →0 √ 2πr σ12 θ=0 [ m 2 N m−2 ] 1 6.5 Mode III When a crack is loaded in Mode III. It can be calculated when more boundary conditions are available in a situation where geometry and load are speciﬁed in more detail.
σ32 = 2µε32 = µu3. it can again be argumented that the exponent λ must be larger than −1. general solution speciﬁc choice ¯ u3 = f + f f = (A + iB)z λ+1 → ¯ f = (A − iB)¯λ+1 z 6. For each value of λ we ﬁnd a ratio between A and B.2 Displacement The general solution of the Laplace equation is written as the sum of a complex function f ¯ and its conjugate f .11 + µu3.22 = 0 → ∇ 2 u3 = 0 6.5.3 Stress components Stress components σ31 and σ32 can now be expressed in the variables r and θ and the constants A.1 . because displacements must remain ﬁnite. These are substituted in the equilibrium equation. ε32 = 2 u3. For a nontrivial solution the equations need to be dependent.2 → equilibrium σ31.1 Laplace equation The two relevant strain components ε31 and ε32 are substututed in Hooke’s law. Choosing a speciﬁc function f with three yet unknown constants A.1 + σ32. σ31 = 2(λ + 1)r λ {A cos(λθ) − B sin(λθ)} σ32 = −2(λ + 1)r λ {A sin(λθ) + B cos(λθ)} σ32 (θ = ±π) = 0 → A B = 0 0 → → 2πλ = nπ → sin(λπ) cos(λπ) sin(λπ) − cos(λπ) det sin(λπ) cos(λπ) sin(λπ) − cos(λπ) = − sin(2πλ) = 0 .5.55 6.2 = µu3. B and λ. The boundary condition that the crack surface is stressfree.2 2 → Hooke’s law σ31 = 2µε31 = µu3. 1 ε31 = 1 u3. B and λ.5.1 . This conﬁnes the values of λ to be a series of discrete numbers. results in two equations for the unknown constants A and B. resulting in two stress components σ31 and σ32 . leading to a Laplace equation for u3 .
6. which is the reason that for the crack tip solution only the ﬁrst term is considered.56 1 n λ = − 2 . . 2 with λ = −1 2 n = 0. III) For the three modes of loading.5 Crack tip solution The SIF KIII is the only unknown constant in the stress tip solution. The only unknown constant B is replaced by the Stress Intensity Factor KIII .6 Crack tip stress (mode I. → .. 2. . . .5. 1. It determines the singularity of the stress ﬁeld. For other crack geometries and loadings. The dependency of geometry is generally indicated with the geometry factor β. the solution of the biharmonic equation results in a series expansion both for the displacement components and the stress components..4 Stress intensity factor The ﬁrst term in the series is the crack tip solution. II. This K solely determines the amplitude of the stresses at the crack tip. equivalent solutions can be derived. √ 2πr σ32 θ=0 KIII = lim r →0 6.5. A=0 → 1 crack tip solution 1 σ31 = Br − 2 {sin( 1 θ)} 2 σ32 = −Br − 2 {cos( 1 θ)} 2 6. It can be determined when more boundary conditions are available. The dependency of the distance r to the crack tip is r λ+1 for the displacements and r λ for the stresses. KIII − sin( 1 θ) σ31 = √ 2 2πr u3 = 2KIII µ r sin( 1 θ) 2 2π KIII 1 cos( 2 θ) σ32 = √ 2πr . The solution is derived for a general crack in a large plate and is summarized below. Each discrete value of the variable λ results in a term of the solution. which only diﬀer in the value of the stress intensity factor K. The ﬁrst term is by far the largest in the vicinity of the crack tip.
as is illustrated in the ﬁgure. KIII σij = √ fIIIij (θ) 2πr √ KIII = βIII τ πa . where the ﬁrst term.7 SIF for speciﬁed cases The SIF can be determined when boundary conditions are speciﬁed in more detail. is the only important one. KII σij = √ fIIij (θ) 2πr √ KII = βII τ πa . The Kzone.1 Kzone The question may arise at what distance to the crack tip. II and III loading. . displacement and stresses are still described accurately by the ﬁrst term of the total solution.6. KI σij = √ fIij (θ) 2πr √ KI = βI σ πa .57 σ σ τ τ τ τ Mode I Mode II Mode III Fig. Some examples for distributed loading are presented. depends on geometry and loading. whose value is determined by the stress intensity factor K. . 6. 6. 6. Solutions for many cases can be found in literature. There is no clear answer to this question.4 : Edge crack under Mode I. 6.5 : Kzone at the crack tip in a threepoint bending test. D Kzone : D I II DII ≪ DI Fig.
8 W 7 2 − 9 2 + 37.6 a W 3 2 1 2 − 5 2 a P/2 S a + 21.7 a W .85 W √ a ≈ 1.36 8.48 W W √ ≈ 1. SIF’s are also known from literature.58 σ 2a W W τ 2a √ πa KI = σ πa sec W √ KII = τ πa 1/2 small a W σ a W √ √ a + KI = σ a 1.48 + W W a 4 53.76 − W a a 2 + 27.7 − 38.12σ πa small W √ √ a KI = σ a 1.12σ πa σ a W a 3 For point loads and a crack loaded by internal pressure.6 W 37.9 W BW 3/2 a 4.12 π − 0.41 W a 2 a 3 18. KI = P W P/2 a PS 2.12 π + 0.
6 1/2 W BW 185. It is called the Fracture Toughness and denoted as Kc .9 W + 655. The local stress growth criterion was formulated in terms of the stress intensity factor K and derived from crack tip stresses. KII = KIIc . We consider a crack in an ”inﬁnite plate” and focus attention on normal stresses and openings in the crack plane. we know which stress is needed and which opening has to be eliminated. Multiplying normal load and closing displacement. This implies that a crack growth criterion can be formulated. First we consider a crack of half length a and then a crack of half length a + ∆a.7 − 2a p W √ KI = p πa p per unit thickness 6. where the stress intensity factor for a certain situation is compared to this critical value. which will be derived in this section. The value of the stress intensity factor has to be calculated.5 a W a 1017 W 3 2 7 2 1 2 − 5 2 9 2 a W P a W a + 638.9 Relation G−K The global crack growth criterion was formulated in terms of the energy release rate G and derived from an energy balance.59 P KI = a P 29. The stress intensity factor determines the ”amplitude” of the crack tip stress for a certain geometry and loading case. To close the latter back to length a again. .8 Kbased crack growth criteria A crack will grow when the crack tip stress exceeds a certain critical value. It may not be surprising that there is a relation between G and K. We may thus conclude that a crack will grow when K reaches a critical value. KIII = KIIIc 6. KI = KIc . The critical value has to be known from experimental measurements.
the energy release rate G. 6. uy = 0 ∆U = 2B a 1 2 σyy a+∆a dx uy = B a σyy uy dx = B f (∆a) ∆a → energy release rate G= 1 B lim ∆a→0 ∆U ∆a = lim f (∆a) = ∆a→0 G= 2 KI E (1 + ν)(κ + 1) 2 (1 + ν)(κ + 1) 2 σ aπ = KI 4E 4E plane stress plane strain G = (1 − ν 2 ) 2 KI E A relation between G and K is also available for multimode loading and can be derived by superposition. This work equals the energy which has been released during opening. crack length a σyy (θ = 0.6 : Central crack opened from length a to length a + ∆a. y x σyy a ∆a Fig. . r = x − a) = crack length a + ∆a σyy (θ = π.60 results in the work needed to accomplish this. r = a + ∆a − x) = 0 √ (1 + ν)(κ + 1) σ a + ∆a √ √ a + ∆a − x uy = E 2 accumulation of elastic energy a+∆a √ σ a 2(x − a) .
10 The critical SIF value In the local crack growth criterion. The actual procedure can be found in literature [25] and is normalized. from which a constant value for KIc results. This can be explained later – we have to consider the plastic crack tip zone for this –.61 G= 1 2 2 2 c1 KI + c2 KII + c3 KIII E plane stress G= 1 2 (K 2 + KII ) E I (1 + ν) 2 (1 − ν 2 ) 2 2 (KI + KII ) + KIII E E plane strain G= 6. the actual value K of the stress intensity factor is compared to a critical value Kc . Dependence of critical Kvalue on the plate thickness B (right).5 KIc σy . It appears that the critical thickness Bc . the critical value would appear to be dependent on the plate thickness. Bc = 2. When KIc would be measured using a large plate with a central crack of length 2a. is very large.7 : Central crack in a plate. σ KI KIc 2a σ Fig. This has to be measured according to an elaborate experimental procedure and is then considered to be a material constant. loaded with Mode I stress σ (left). 6. This means that such an experimental setup would be unfeasible and therefore other experiments are designed and normalized. 2 B Bc B √ KIc = σc πa .
carbon Al Al Al Al 2014T4 2024T3 7075T651 7079T651 σv [MPa] 1669 2241 1496 1827 345 241 448 393 545 469 1103 1083 945 √ KIc [MPa m ] 93.5 66.7 33.62 6.4 38. D6AC steel.0 47.3 197.1 29.5 37.8 219. They are taken from [25]. Material steel. 300 maraging steel.1 KIc values The table lists values for the fracture toughness KIc for various materials. 350 maraging steel. AISI 4340 steel.10.8 28.0 38.6 34.5Si . A533B reactor steel.4 70.3 Ti 6Al4V Ti 6Al6V2Sn Ti 4Al4Mo2Sn0.
1 Stress component transformation We consider a plane with unity normal vector n and stress vector p. The crack tip stresses can be determined as the superposition of the stress components due to separate Mode I and Mode II loadings.Chapter 7 Multimode crack loading In practical situations.r. e2 } are stored in columns n and p. Mode I Mode II Mode I + II Mode I + II Fig. A tensile load parallel to the crack can be added because it has no inﬂuence on the crack tip stress. Mode II and a combination of Mode I and II.r.t. In the ﬁgure the load is visualized as an edge load on a square material volume element with edges parallel and perpendicular to the crack.t. a crack is mostly subjected to a combined Mode I and Mode II loading. e2 }. Components 63 . this is rotated anticlockwise over an angle θ w. 7.1 : Central crack loaded in Mode I. Each random load can be transformed to this loading situation by proper transformation of the edge loads. e2 } ˜ ˜ of n and p w. 7. Components of n and ∗ ∗ p with respect to the base {e1 . {e1 . Another base {e1 .
e2 }.r. the basis {e ∗ ˜e ∗ } can be calculated from σ. matrix T . a ˜ be determined which relates the columns n∗ to n. Using the relation between {e1 . and p∗ to p. 2 p = σn → ˜ ∗ ˜ T p = σ T n∗ → p∗ = T T σ T n∗ = σ ∗ n∗ ˜ ˜ ˜ ˜ ˜ σ∗ = T T σ T → σ = T σ∗T T ∗ ∗ σ11 σ12 ∗ ∗ σ21 σ22 → = = c s −s c σ11 σ12 σ21 σ22 cσ11 + sσ12 cσ21 + sσ22 c −s s c −sσ11 + cσ12 −sσ21 + cσ22 c2 σ11 + 2csσ12 + s2 σ22 − csσ11 + (c2 − s2 )σ12 + csσ22 = −csσ11 + (c2 − s2 )σ12 + csσ22 2 σ − 2csσ + c2 σ s 11 12 22 c s −s c . the matrix σ 1. the stress vector components p can be calculated from the ˜ Cauchy stress matrix σ and the components n of the normal vector. ∗ e1 = cos(θ)e1 + sin(θ)e2 = ce1 + se2 ∗ e2 = − sin(θ)e1 + cos(θ)e2 = −se1 + ce2 ∗ ∗ p = p1 e1 + p2 e2 = p∗ e1 + p∗ e2 → 1 2 c −s p1 p∗ 1 → = p∗ p2 s c 2 p∗ 1 p∗ 2 idem = : c s −s c n∗ = T T n ˜ ˜ p1 p2 → p = T p∗ ˜ ˜ → p∗ = T T p ˜ ˜ With respect to the basis {e1 . Using the transformation ∗ w.2 : Material plane with unit normal vector and stress vector.64 ∗ ∗ base are stored in columns n∗ and p∗ . e2 }. ˜ transformation matrix T can ˜ ˜ ˜ ˜ (b) p e2 ∗ e2 n ∗ e1 θ e1 Fig. 7.t. e2 } and {e1 .
With the ﬁgure below. it is easily understood that cylindrical stress components can be derived from Cartesian components by transformation of the stress matrix.65 Stress component transformation can also be used to determine stress components in a cylindrical coordinate system. .4 : Cartesian stress cube at the crack tip. σyy σxy et σxx er r e2 e1 θ σrt σrr σtt Fig. 7. These will be needed later in crack growth direction criteria. er = ce1 + se2 σrr σrt = σtr σtt . c s −s c et = −se1 + ce2 σxx σxy c −s σxy σyy s c c2 σxx + 2csσxy + s2 σyy − csσxx + (c2 − s2 )σxy + csσyy = −csσxx + (c2 − s2 )σxy + csσyy 2σ 2σ s xx − 2csσxy + c yy Crack tip stresses : Cartesian The Cartesian crack tip stresses have been derived before and are summarized below. σyy σxy σxx Fig. 7.3 : Cartesian and cylindrical stress ’cubes’ at the crack tip.
σrr σrt θ σtt Fig. 7. The functions frr . ftt and frt can be derived.5 : Cylindrical stress ’cube’ at the crack tip.66 KII KI fIxx (θ) + √ fIIxx (θ) σxx = √ 2πr 2πr KI KII σyy = √ fIyy (θ) + √ fIIyy (θ) 2πr 2πr KI KII σxy = √ fIxy (θ) + √ fIIxy (θ) 2πr 2πr θ fIxx(θ) = cos( 2 ) 1 − sin( θ ) sin( 3θ ) 2 2 θ fIyy (θ) = cos( 2 ) 1 + sin( θ ) sin( 3θ ) 2 2 θ θ fIIyy (θ) = sin( 2 ) cos( 2 ) cos( 3θ ) 2 θ fIxy (θ) = sin( θ ) cos( 2 ) cos( 3θ ) 2 2 θ θ fIIxx(θ) = − sin( 2 ) 2 + cos( 2 ) cos( 3θ ) 2 θ fIIxy (θ) = cos( 2 ) 1 − sin( θ ) sin( 3θ ) 2 2 Substitution in the transformation relation. KII KI fIrr (θ) + √ fIIrr (θ) σrr = √ 2πr 2πr KI KII σtt = √ fItt (θ) + √ fIItt (θ) 2πr 2πr KI KII σrt = √ fIrt (θ) + √ fIIrt(θ) 2πr 2πr fIrr (θ) = fItt (θ) = 5 4 3 4 1 cos( θ ) − 4 cos( 3θ ) 2 2 3 5 fIIrr (θ) = − 4 sin( θ ) + 4 sin( 3θ ) 2 2 θ cos( 2 ) + 1 cos( 3θ ) 4 2 3 4 3 fIItt (θ) = − 4 sin( θ ) − 2 1 4 sin( 3θ ) 2 fIrt (θ) = sin( θ ) + 2 1 4 sin( 3θ ) 2 fIIrt (θ) = 1 4 θ cos( 2 ) + 3 cos( 3θ ) 4 2 . using some trigonometry. results in cylindrical crack tip stresses.
load ∗ σ22 ∗ σ12 ∗ σ12 2a ∗ σ12 ∗ σ11 ∗ σ22 c2 σ11 + 2csσ12 + s2 σ22 ∗ ∗ σ11 σ12 − csσ11 + (c2 − s2 )σ12 + csσ22 = ∗ ∗ 2 − s2 )σ + csσ −csσ11 + (c σ21 σ22 12 22 2 σ − 2csσ + c2 σ s 11 12 22 KII KI fIij (θ) + √ fIIij (θ) crack tip stresses sij = √ 2πr 2πr ∗ √ ∗ √ with KI = β σ22 πa . 7.t. Note that in the next cases the load is indicated by σij and the crack tip stresses by sij . KII = γ σ12 πa Example multimode load As an example we consider the biaxial load on a crack which orientation has an angle θ with the horizontal direction.2 Multimode load For a crack with angle θ w. The load parallel to the crack (σ11 ) is assumed to have no inﬂuence on the crack tip stress state. σ22 σ12 2a σ11 e2 e2∗ e1∗ θ e1 Fig.7 : Transformation of multimode load into Mode I and II loads. σ 2a kσ θ ∗ σ22 ∗ σ12 ∗ σ11 ∗ σ12 2a Fig.67 7. the edge of the material volume element with edge loads represented by σ. The crack tip stresses of the combined loading are ∗ determined by superposition.6 : Transformation of multimode load into Mode I and II loads.r. . 7. the Mode I and Mode II loading is represented by the edge loads σ ∗ on a rotated material volume element.
σt = pR pR 1 1 =σ .r. The crack has an angle θ w. 7. σ12 = cs(1 − 1 )σ = 2 cs σ σ22 = s2 1 σ + c2 σ 2 2 √ √ pR √ ∗ 1 KI = σ22 πa = ( 2 s2 + c2 )σ πa = ( 1 s2 + c2 ) πa 2 t pR √ ∗ √ 1 1 πa KII = σ12 πa = 2 cs σ = 2 cs t .t. the axial direction. σt 2a σa ∗ σ22 ∗ σ12 p R t θ ∗ σ11 Fig. can be used directly in the situation of a crack through the whole thickness of the wall of a thinwalled tube.68 load ∗ σ11 = c2 σ11 + 2csσ12 + s2 σ22 = c2 kσ + s2 σ ∗ σ12 = −csσ11 + (c2 − s2 )σ12 + csσ22 = cs(1 − k)σ ∗ σ22 = s2 σ11 − 2csσ12 + c2 σ22 = s2 kσ + c2 σ crack tip stresses KII KI fIij (θ) + √ fIIij (θ) sij = √ 2πr 2πr stress intensity factors √ ∗ √ KI = βI σ22 πa = βI (s2 k + c2 )σ πa √ ∗ √ KII = βII σ12 πa = βII cs(1 − k)σ πa Example multimode load The results from the example above.8 : Crack in tube wall at an angle θ with the tube axis. σa = = 2σ → k= t 2t 2 ∗ ∗ 1 .
For both criteria.t. The hypothesis is that a crack tip will move in the direction of the point where the highest value of the tangential stress component is found. in multimode loading. 7. we have to use stress and displacement components in a cylindrical coordinate system. When the crack is loaded only in mode I.9 : Cylindrical stress components at the crack tip.69 7. Two crack growth direction criteria will be described in the following : • maximum tangential stress (MTS) criterion. 7. More general cases are explored with numerical techniques.r.3 Crack growth direction Crack growth criteria are used to check whether a crack will grow at a given external load. However. it will be expected that the crack tip will move in the crack plane. Hypothesis : crack growth towards local maximum of σtt ∂ 2 σtt <0 ∂θ 2 → ∂σtt =0 ∂θ and → crack growth θc KIc σtt (θ = θc ) = σtt (θ = 0) = √ 2πr . This can be expressed mathematically by ﬁrst and second derivatives of the tangential stress component w. the angle θ.3. the crack growth direction is not so easy to predict. σrr σrt θ σtt Fig.1 Maximum tangential stress criterion The ﬁrst crack growth direction criterion to be discussed here is the maximum tangential stress criterion. For special loading cases it will be possible to calculate the crack growth direction by analytical means. which was suggested and published by Erdogan and Sih in 1963 [18]. • strain energy density (SED) criterion.
The requirement that the ﬁrst derivative of σtt (θ) is zero.70 Because the function σtt (θ) is known. ∂σtt = KII (3 cos(θc ) − 1) = 0 → θc = ± arccos( 1 ) = ±70. the relations can be elaborated. leads to an equation. ∂σtt =0 → ∂θ K KII θ θ 3√ I − 1 sin( 2 ) − 1 sin( 3θ ) + 3 √ − 1 cos( 2 ) − 2 4 4 2 2 4 2πr 2πr KI sin(θ) + KII {3 cos(θ) − 1} = 0 ∂ 2 σtt <0 → ∂θ 2 θ 3 KII 3 KI √ − 1 cos( 2 ) − 3 cos( 3θ ) + 4 √ 4 4 4 2 2πr 2πr 3 4 cos( 3θ ) = 0 2 → 1 4 θ sin( 2 ) + 9 sin( 3θ ) < 0 4 2 KIc σtt (θ = θc ) = √ → 2πr 1 KI 1 KII c 3 cos( θ2c ) + cos( 3θc ) + 4 −3 sin( θ2 ) − 3 sin( 3θc ) = 1 4K 2 2 KIc Ic Mode I load For mode I loading of the crack. so that we can conclude that the crack tip will move in that direction. Only for the negative value θc = −70. it is expected that the crack tip will advance in the plane of the crack.6o ∂θ 2 θc KIc σtt (θc ) = √ → KIIc = 3 KIc 4 2πr .6o 3 ∂θ 2σ ∂ tt <0 → θc = −70.6o . the second requirement is satisﬁed. which results in two candidate values for θc . Indeed this obviously results from solving the relations for θ. the crack growth direction has to be determined from the relations for θ. The critical value of the mode II stress intensity factor can be expressed in the critical value of the mode I SIF. ∂σtt = KI sin(θ) = 0 ∂θ ∂ 2 σtt <0 ∂θ 2 θc KIc σtt (θc ) = √ → 2πr → θc = 0 KI = KIc Mode II load For pure mode II loading of the crack.
6 0.3 0.8 1 K /K I Ic K /K I Ic Fig. KI [− sin( θ ) − sin( 3θ )] + KII [− cos( θ ) − 3 cos( 3θ )] = 0 2 2 2 2 KI [3 cos( θ ) + cos( 3θ )] + KII [−3 sin( θ ) − 3 sin( 3θ )] = 4KIc 2 2 2 2 −KI f1 − KII f2 = 0 −KI f2 + KII f3 < 0 KI f4 − 3KII f1 = 4KIc KII KI f1 − f2 = 0 − KIc KIc KI KII − f2 + f3 < 0 KIc KIc KII KI f4 − 3 f1 = 4 KIc KIc KI [− cos( θ ) − 3 cos( 3θ )] + KII [sin( θ ) + 9 sin( 3θ )] < 0 2 2 2 2 → KI For a range of values 0 ≤ KIc ≤ 1 a range of values for the crack growth direction angle 0 ≤ θ ≤ 90o is evaluated to decide if the relations are satisﬁed.11 : Crack growth direction and Mode ratio for multimode loading.8 1 −30 θc −40 −50 −60 −70 0 0.6 0. Multimode load Determining the crack growth direction for multimode loading is only possible using numerical calculations. The relations which have to be satisﬁed for crack growth in a certain KI direction are written is terms of the parameter KIc .4 0. In the ﬁrst one KII is plotted KIc KI as a function of KIc and it can be used to determine which combination leads to crack growth. 7.1 0 0 0.8 0.2 0. 0 −10 −20 Ic II 0.4 0.71 τ θc τ Fig.2 0. . From the second plot the crack growth angle θc can then be determined.10 : Crack growth direction for Mode II loading.5 0.9 0.7 0.4 0.2 0.6 K /K 0. 7.
θ) Hypothesis : crack growth towards local minimum of SED ∂2S >0 ∂θ 2 ∂S =0 ∂θ and → → θc crack growth S(θ = θc ) = S(θ = 0. 1 ν 1 2 2 2 2 2 (σ 2 + σyy + σzz ) − (σxx σyy + σyy σzz + σzz σxx ) + (σ + σyz + σzx ) 2E xx E 2G xy Ui = KII KI θ θ cos( θ ) 1 − sin( 2 ) sin( 3θ ) − √ sin( 2 ) 2 + cos( θ ) cos( 3θ ) σxx = √ 2 2 2 2 2πr 2πr KI KII θ θ σyy = √ cos( θ ) 1 + sin( 2 ) sin( 3θ ) + √ sin( 2 ) cos( θ ) cos( 3θ ) 2 2 2 2 2πr 2πr KI KII θ θ σxy = √ sin( θ ) cos( 2 ) cos( 3θ ) + √ cos( 2 ) 1 − sin( θ ) sin( 3θ ) 2 2 2 2 2πr 2πr . the crack tip will move toward the point where S is minimum. The strain energy density criterion is proposed and published by Sih in 1973 [61] and states that at crack growth. pl. The strain energy density factor is than also known and dependent on θ. the strain energy density function can be expressed in the stress components and the material parameters E and ν. KI and KII . The requirements for the crack growth direction result in two relations. εij Ui = Strain Energy Density (Function) = 0 σij dεij S = Strain Energy Density Factor = rUi = S(KI . θ.72 7. from which θc can be determined.2 Strain energy density (SED) criterion The strain energy density Ui is the stored elastic energy per unit of volume.strain) = Sc For the case of linear elastic material behavior. For linear elastic material behavior this speciﬁc energy is easily calculated in the crack tip region and appears to be inverse proportional to the distance to the crack tip. KI and KII . At the crack tip these stress components are known functions of r.3. as such being independent of r. The strain energy density factor S is now deﬁned as the product rUi . KII .
which is the value for plane strain.73 2 2 S = rUi = S(KI . The obvious crack growth direction is θc = 0. The two requirements lead to two possible crack growth directions. pl. From this value the critical shear stress can be derived. The result diﬀers for plane stress and plane strain.strain) = KIc 2πE Mode II load √ For Mode II loading we have KI = 0 and KII = τ πa. The ﬁrst requirement results in two possible values of θc . leading to κ = 3. Crack growth will take place if the value of S(θc ) equals the critical value Sc . KII . For ν = 0. the direction from SED is the same as the one from MTS. σ2 a {1 + cos(θ)}{κ − cos(θ)} 16G → arccos → 1 2 (κ 2 S = a11 kI = ∂S = sin(θ){2 cos(θ) − (κ − 1)} = 0 ∂θ θc = 0 or ∂2S ∂θ 2 = 2 cos(2θ) − (κ − 1) cos(θ) > 0 − 1) θc = 0 S(θc ) = σ2a σ2a {2}{κ − 1} = (κ − 1) 16G 8G (1 + ν)(1 − 2ν) 2 Sc = S(θc . Crack growth will occur if the minimum Svalue equals the critical value Sc . because 1 ≤ κ ≤ 3. θ) = a11 kI + 2a12 kI kII + a22 kII with a11 a12 a22 ki 1 = 16G (1 + cos(θ))(κ − cos(θ)) 1 = 16G sin(θ){2 cos(θ) − (κ − 1)} 1 = 16G {(κ + 1)(1 − cos(θ)) + (1 + cos(θ))(3 cos(θ) − 1)} √ = Ki / π Mode I load √ For Mode I loading we have KII = 0 and KI = σ πa. which was determined in pure Mode I loading. 2 S = a22 kII = S(θc ) = τ 2a 1 { (−κ2 + 14κ − 1)} 16G 12 192GSc 1 S(θc ) = Sc → τc = √ 2 + 14κ − 1 a −κ τ 2a [(κ + 1){1 − cos(θ)} + {1 + cos(θ)}{3 cos(θ) − 1}] 16G ∂S = sin(θ) [−6 cos(θ) + (κ − 1)] = 0 ∂θ → ∂2S 2 = 6 − cos (θ) + (κ − 1) cos(θ) > 0 ∂θ 2 θc = ± arccos 1 (κ − 1) 6 . The second requirement is only satisﬁed for θc = 0.
4 ν=0.75 ν=0.2 ν=0.3 ν=0.75 ν=0.495 1 0 0.4 0.7 −30 0.1 −90 0 0 0.6 K /K Ic II θc −40 0.4 −60 0.t.2 −80 0.1 ν=0.0.8 0.6 0.74 Multimode load It is again not possible to determine the crack growth direction for a general multimode loading.75 ν=0. The requirements according to the SED criterion can be used to test a sequence of θc values for a sequence of β values and a plot can be made.1 0.3 0.7 0.3 ν=0. √ kII = σ a sin(β) cos(β) S = σ 2 a sin2 (β) a11 sin2 (β) + 2a12 sin(β) cos(β) + a22 cos2 (β) ∂S = (κ − 1) sin(θc − 2β) − 2 sin{2(θc − β)} − sin(2θc ) = 0 ∂θ ∂2S = (κ − 1) cos(θc − 2β) − 4 cos{2(θc − β)} − 2 cos(2θc ) > 0 ∂θ 2 .1 0.2 0.8 −20 0. 7. √ kI = σ a sin2 (β) .7 0.75 ν=0.4 −60 0.5 KI/KIc 0.7 0.5 ν=0.6 0.5 −50 0.3 0.3 0. For this situation the scaled stress intensity factors kI and kII can be calculated.9 1 0 0 0.3 0.1 ν=0.8 0.5 KI/KIc 0.2 ν=0.0.4 0.8 −20 0.0.2 0.6 0.2 ν=0.2 0. The ﬁgure below shows a crack oriented at an angle β w.6 K /K 0 0.7 ν=0.495 −30 0.9 1 Ic II θc −40 0.13 : Crack growth direction and Mode ratio for multimode loading.1 0. however.1 ν=0.2 −80 0.3 −70 0. The ﬁgure shows the result for plane strain.5 K /K I 0.2 ν=0.1 −90 0 0. 10 ν=0.1 ν=0.12 : Crack growth direction and Mode ratio for multimode loading. the tensile load σ.9 −10 0.7 0.1 0.9 −10 0.5 K /K I 0.8 0. 10 ν=0. Numerically it can be done.9 1 Ic Fig.3 ν=0. 7.r.4 ν=0.3 ν=0.4 0.495 −50 0.495 1 0 0.2 0.4 0.4 ν=0. from which follows the expression for S as a function of β and θ.0.4 ν=0.6 0.8 0.9 1 Ic Fig.3 −70 0.
75
σ
−θc
β 2a
θc
90 80 70 60 50 40 30 20 10 0
ν = 0.5
ν=0 ν = 0.1 0 10 20 30 40 50 60 70 80 90 β
σ
Fig. 7.14 : Crack growth directions for multimode loading.
76
Chapter 8
Dynamic fracture mechanics
When a crack is subjected to impact loading, the crack tip stresses and strains can only be calculated by taking into account dynamic eﬀects like material acceleration and solving the equations of motion. This is the subject of Dynamic Fracture Mechanics (DFM) and understandably involves rather sophisticated mathematics. Another subject, studied in DFM, is the behavior of a crack, which grows at high rate under (quasi)static loading after reaching its critical length. In this chapter an approximate value for the crack tip speed is derived, using the global energy balance, taking into account the kinetic energy term. This approach is referred to as the kinetic approach. In DFM a socalled static approach is also used to study the behavior of cracks, subjected to thermal shock loading.
8.1
Crack growth rate
To calculate the crack tip speed or crack growth rate, the general energy balance is used as a starting point. It is elaborated for a crack with length 2a in an inﬁnitely large plate with uniform and constant thickness B. The crack is subjected to a Mode I load σ, resulting from an external load applied at edges far away from the crack. This implies that dUe = 0 when the crack length changes. It is assumed that the presence of the crack has led to a complete release of elastic energy Ui in the elliptical area, indicated in the ﬁgure. Dissipation is assumed to be zero, so all available energy is going to be transformed into surface energy Ua and kinetic energy Uk .
dUa dUd dUk dUe dUi − = + + da da da da da
77
8. When the crack opens. Its value is indicated as ak(a) and will be considered later. However. here assumed to be in a state of plane stress. it is ˙ assumed that the crack tip speed s is independent of the (half) crack length a. → σ2 E → dUd =0 da dUa = 4γB da − dUi 2πaBσ 2 = da E Kinetic energy The total kinetic energy of the plate is the integral over its volume of the speciﬁc kinetic ˙ ˙ ˙y ˙x energy 1 ρ(u2 + u2 ). where ρ is the material density and ux and uy the material velocity 2 components. To allow further derivations.78 σ y thickness B 2a a x σ Fig. 1 Uk = 2 ρB Ω (u2 + u2 ) dxdy ˙x ˙y u x ≪ uy = ˙ ˙ duy da 2 material velocity Uk = 2 1 2 ρs B dxdy ds =0 da Ω assumption duy da duy duy = = s dt da dt da . The kinetic energy change at crack growth can now be calculated by integration. dUe =0 da Ua = 4aBγ 1 Ui = 2πa2 B 2 . the material velocity perpendicular to the crack faces (uy ) is ˙ much higher than the velocity in the crack plane (ux ).1 : Central crack in large plate with unloaded elliptical area. the integral cannot be calculated analytically. The displacement uy of the crack face (θ = π) is available from the analytical solution for a static crack.
34 3670 aluminum 70 2700 0. The measured crack tip speed s is also listed. This leads to the ﬁnal approximation for the crack tip speed. the crack will grow at high speed. so that the actual (half) crack length a will be much larger than the critical crack length ac .51. The ratio s/c takes values between 0. As can be seen from the expression for s.25 5300 2000 0.38 c 1 − a a ≫ ac 2γE ac = πσ 2 .4. so the result must be an approximation. The earlier assumption that s is independent of a does not appear to hold. steel E [GPa] ρ [kg/m2 ] ν c [m/sec] s [m/sec] s/c 210 7800 0. density ρ and calculated sound speed c. Roberts and Wells did this in 1954 [56] and came up with k = 43.38 rubber 20 900 0. Mostly the experimental ratio is found to be smaller.38 is well within this range.38 k ac s = 0. More approximation is involved in the calculation of k(a).2 and 0.79 d duy 2 dUk 1 = 2 ρs2 B dxdy da da Ω da √ σ 2a − x √ σ duy a2 − ax → = 2 √ uy = 2 2 E da E a2 − ax σ E 2 2 dUk = ρs2 B da a Ω 1 x2 (x − 2a) dxdy = ρs2 B a3 (a − x)2 σ E a k(a) Substitution of all results in the energy balance leads to an expression for the crack tip speed s where the critical crack length ac and the sound speed c can be introduced. c= E ρ → s ≈ 0.5 46 . due to the dissipation during crack growth. σ 2πaσ 2 = 4γ + ρs2 E E s= E ρ 1 2 2 ak 1 2 → 2π k 1 2 2γE 1− πaσ 2 .34 5090 glass 70 2500 0. which is done numerically.29 5190 1500 0. 1 2 2π ≈ 0.38 c For some materials the table lists values for Young’s modulus E. so the approximation 0.29 copper 120 8900 0.
3 Crack tip stress When a crack tip moves at high speed through a material. E ρ κ+1 κ−1 µ ρ C0 C1 C2 = elongational wave speed = = dilatational wave speed = shear wave speed = = µ ρ CR = Rayleigh velocity = 0. e. just as in the static case. respectively. the elastic stresses σDij at the crack tip can be calculated [70].r.38 C0 1 − s = CR 1 − ac a ac a 8. r. the crack plane. with wave front speeds C1 and C2 .62 C0 Corrections More elaborate derivations of the crack tip speed have been published. however. dilatation (volume change) and shear. In DFM more sophisticated analyzes of crack growth rate use other deformation modes. also indicated as C0 .t. The ﬁrst term in this stress ﬁeld appears to be proportional √ to KD / 2πr. In this dynamic case. the sound speed c was introduced. a function of the angle θ w. Dulancy & Brace (1960) Freund (1972) s = 0.2 Elastic wave speeds In the derivation of the crack tip speed s. s. KD fij (θ. with r the distance to the crack tip and KD the Dynamic Stress Intensity Factor.2 < s < 0. E. ν) σDij = √ 2πr .g. Frequently used is the Rayleigh velocity CR of wave fronts traveling at the surface of a material. The stress components are. Dulancy & Brace (1960) derived an expression without assuming s to be independent of a.54 C0 a ´ 0.80 0. which is the speed of a wave front with elastic elongational strain. they also depend on material parameters and the crack tip speed.4 c 8. Freund (1972) related s to the Rayleigh velocity CR .
The ratio of the tangential stress σIDtt (θ) and the tangential stress in the crack plane (θ = 0). and will soon lead to complete material failure. where it is subjected to diﬀerent loading conditions. r. which may result in crack arrest. illustrating crack branching. the crack will move out of the crack plane.2 Fast fracture and crack arrest A crack.87 θ crack branching volgens MTS Fig. ν) σDij = √ 2πr 0. can reach a location.87 the curve has a maximum value at θ ≈ π/4. This crack branching will occur repeatedly. 8. being the minimal value of a measured Dynamic Fracture Toughness KDc .6 0 π π 2 s cR σtt 0. The graphical representation shows diﬀerent curves for various values of the crack tip speed s. KID fij (θ. 8. which runs through a material.2 : Tangential stress ratio as a function of crack growth direction. T ) KD < min 0<s<CR KDc (s.3. For s/CR ≈ 0. E. According to the Maximum Tangential Stress criterion. KD ≥ KDc (s.1 Crack branching Cylindrical stress components for Mode I loading can be derived by transformation as has been done for the static loading case. This will occur when KD is below a critical value KA . when the crack runs through the material. which is a function of crack tip speed s and temperature T . s.81 8. T ) = KA → → crack growth crack arrest .9 σDtt (θ) σDtt (θ = 0) 1 max 0. can be calculated as a function of θ.3.
where a crack is forced to run from a low temperature region into a higher temperature region. the Robertson Crack Arrest Temperature (CAT) test is done. using instrumented DCB specimens.3 : Instrumented DCB specimen to measure crack propagation speed [39]. sometimes in combination with high speed photography (typically 106 frames/sec). 8.4 Experiments In DFM experiments are done to measure crack tip speeds.82 8. To determine the temperature dependent KDc . Fig. .
In threedimensional cases. the occurrence of yielding is tested by means of a yield criterion. The Von Mises yield criterion can be expressed in the principal stresses. Both are formulated in terms of principal stresses σ1 . σ2 and σ3 . For plane stress cases σ3 = 0 and for plane strain situations Hooke’s law gives σ3 = ν(σ1 + σ2 ). we employ the Von Mises and the Tresca criteria. The critical value is determined from a tensile test and expressed in the yield stress σy . 9. In reality the material will yield before the crack tip is reached and the elastic solution is no longer valid. with ν as Poisson’s ratio. Von Mises W d = Wcd 2 (σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 = 2σy τmax = τmaxc σmax − σmin = σy Tresca 83 . where the ﬁrst two are in the plane of the plate and σ3 is in the normal direction. In this chapter the region at the crack tip where yielding occurs is determined. Tresca yield criterion is based on the hypothesis that yielding occurs when the maximum shear stress τmax reaches a critical value τmaxc . Postyield hardening is not considered. Here. The Tresca yield criterion can be expressed in the diﬀerence of the maximum and the minimum principal stresses. This critical value is again determined in a tensile test and related to the yield stress σy . so the material behaves ideal plastic. The inﬂuence of this plastic crack tip zone on the crack behavior is explained.Chapter 9 Plastic crack tip zone The elastic crack tip stresses reach an inﬁnite value. when the distance to the crack tip decreases to zero.1 Von Mises and Tresca yield criteria Von Mises yield criterion is based on the hypothesis that yielding occurs when the speciﬁc distorsional elastic energy W d reaches a critical value Wcd .
because the value of the third principal stress depends on Poisson’s ratio. For plane strain the smallest value may be the second principal stress. crack tip stresses KI fIij (θ) σij = √ 2πr KI θ σ1(+). For plane strain the third principal stress is not zero but related to the ﬁrst two principal stresses.84 9.2 Principal stresses at the crack tip For both the Von Mises and the Tresca yield criterion. plane stress state σxx σxy 0 σ = σxy σyy 0 0 0 0 σzz = σzx = σzy = 0 → det(σ − σI) = 0 → 1/2 1/2 → 2 σ σ 2 − σ(σxx + σyy ) + (σxx σyy − σxy ) = 0 1 4 (σxx 1 4 (σxx 2 − σyy )2 + σxy 2 − σyy )2 + σxy σ1 = 1 (σxx + σyy ) + 2 σ2 = 1 (σxx + σyy ) − 2 σ3 = 0 plane strain state σ3 = ν(σ1 + σ2 ) Elastic crack tip stress components for Mode I are used to derive the principal crack tip stresses. KI σ2 = √ cos( θ ){1 − sin( θ )} 2 2 2πr 2νKI cos( θ ) σ3 = √ 2 2πr plane stress plane strain σ 1 > σ2 > σ3 σ 1 > σ2 > σ3 or σ 1 > σ3 > σ2 .2(−) = √ cos( 2 )± 2πr 1 4 θ −2 cos( 2 ) sin( θ ) sin( 3θ ) 2 2 2 θ + sin( θ ) cos( 2 ) cos( 3θ ) 2 2 2 KI θ σ1 = √ cos( θ ){1 + sin( 2 )} 2 2πr σ3 = 0 or . They are the eigenvalues of the Cauchy stress matrix and for the plane stress case they can be calculated analytically. For plane stress the third principal stress is zero and the smallest. the principal stresses have to be used. The ﬁrst principal stress is always the largest.
25 1000 800 600 σ 400 200 0 σ 1000 800 600 400 200 0 ν = 0.35 0 20 40 θ 60 80 100 0 20 40 θ 60 80 100 ν = 0.3 Von Mises plastic zone In the Von Mises yield criterion all principal stresses are used.85 ν = 0. 9. The third principal stress depends on Poisson’s ratio. The derivation uses the next trigonometric relations : θ θ cos(θ) = cos2 ( θ ) − sin2 ( 2 ) = 2 cos2 ( 2 ) − 1 2 θ cos2 ( θ ) sin2 ( 2 ) 2 = 1 4 sin (θ) 2 → cos2 ( θ ) = 2 1 2 {1 + cos(θ)} Von Mises yield criterion 2 (σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 = 2σy plane stress σ3 = 0 2 2 2 (σ1 − σ2 )2 + σ2 + σ1 = 2σy 2 KI 2 cos2 ( θ ) 6 sin2 ( θ ) + 2 = 2σy 2 2 2πrv 2 2 KI KI θ ry = 1 + cos(θ) + cos2 ( 2 ) 1 + 3 sin2 ( θ ) = 2 2 2 2πσy 4πσy 3 2 sin2 (θ) plane strain σ3 = ν(σ1 + σ2 ) .5 0 20 40 θ 60 80 100 0 20 40 θ 60 80 100 Fig.1 : Principal stresses for plane strain as a function of the direction angle.45 1000 800 600 σ 400 200 0 σ 1000 800 600 400 200 0 ν = 0. 9. For plane stress and plane strain the distance ry of the elastic/plastic boundary is easily calculated.
5 1 1. Von Mises plastic zones 1 pl. It appears that the plane stress plastic zone is larger than the plane strain plastic zone.86 2 KI 2 cos2 ( θ ) 6 sin2 ( θ ) + 2(1 − 2ν)2 = 2σy 2 2 2πry 2 KI (1 − 2ν)2 {1 + cos(θ)} + 3 sin2 (θ) ry = 2 2 4πσy 2 2 2 (ν 2 − ν + 1)(σ1 + σ2 ) + (2ν 2 − 2ν − 1)σ1 σ2 = σy The elastic/plastic boundaries can be plotted in an xycoordinate system with the crack tip in the origin and the crack along the line −∞ < x < 0. however.5 0 −0. depending on the coordinate angle θ and on Poisson’s ratio ν.4 Tresca plastic zone Using the Tresca yield criterion implies using the maximum and minimum principal stresses. σmin } = {σ1 .5 0 0. 9.5 −1 −0. the minimal principal stress can be σ3 or σ2 . For the plane strain case. 9. For the plane stress case σ3 = 0 is always the minimum.2 : Plastic zones according to the Von Mises yield criterion.stress pl.strain 0. σmax − σmin = σy plane stress {σmax . y = 0.5 Fig. For plane strain two diﬀerent elastic/plastic boundaries can be calculated. σ3 } .
when . σ3 } 2 ry = plane strain II 2 KI θ θ (1 − 2ν) cos( 2 ) + cos( 2 ) sin( θ ) 2 2 2πσy σ 1 > σ3 > σ2 → {σmax . σmin } = {σ1 . the plastic region is deﬁned by two curves.strain sig3 = min pl.3 : Plastic zones according to the Tresca yield criterion. one for σ1 > σ2 > σ3 (smallest) and one for σ1 > σ3 > σ2 (largest).5 Fig.5 0 −0. 9.stress pl. In a thin plate the material can contract freely perpendicular to the plate (3direction).5 1 1. Tresca plastic zones 1 pl.87 KI θ cos( 2 ) + cos( θ ) sin( θ ) 2 2 2πry ry = plane strain I = σy 2 2 KI cos( θ ) + cos( θ ) sin( θ ) 2 2 2 2 2πσy σ 1 > σ2 > σ3 → {σmax .5 0 0.strain sig2 = min 0. 9.5 Inﬂuence of the plate thickness The plastic crack tip zone is considerably diﬀerent for a crack in a thin and a thick plate. For the plane strain situation. σ2 } ry = 2 KI sin2 (θ) 2 2πσy Plotting the elasticplastic boundary reveals again that the largest plastic zone occurs for the plane stress case. σmin } = {σ1 .5 −1 −0.
the shear planes can be visualized in some cutting planes.88 an inplane load is applied. which is at a much lower stress and has no intention at all to contract. For the thick plate. Employing an etching procedure. A typical dogbone plastic zone develops at the crack tip. The crack tip material at greater depth from the surface is prohibited to contract by the surrounding material. which is empirically related to KIc and σy . 9. Fig. which are perpendicular. The inner crack tip material is thus in a state of plane strain.4 : Plastic crack tip zone for a thin and a thick plate.6 Shear planes Plastic deformation takes place by sliding of crystallographic planes induced by shear stresses. the value is lower and becomes independent of the thickness. only the surface layers are free to contract in normal direction and are thus in a plane stress state. when this is larger than a threshold value Bc . This explains that measured KIc values are higher when determined with thin test specimens. . The plane stress plastic crack tip zone develops over the whole plate thickness.5 KIc σy 2 9. For thicker specimens. In the surface layers we ﬁnd plane stress plastic zones and in the deeper bulk material the plastic crack tip zone is smaller as it is associated with plane strain. Critical plate thickness In a thin plate the relative volume of yielding material is much larger than in the thick plate. Over the whole plate thickness a plane stress state exists with σ3 = 0. Bc > 25 3π KIc σy 2 > 2. The planes with the maximum shear stresses can be found at an angle of 45o between the planes with the maximum and minimum principal stresses.
5 0 −0. can also be used as a multiplication factor of the yield stress to take into account other principal stresses.5 2 y Fig.5 : Shear planes at the crack tip: plane stress (left). 9. 9. The pcf. 1 2 {(σ1 − σ2 )2 + (σ2 − σ3 )2 + (σ3 − σ1 )2 } = 1 σmax =√ σy 1 − n − m + n2 + m2 − mn 1 − n − m + n2 + m2 − mn σmax = σy pcf = In a plane stress state with m = 0.5 −1 −1 σ /σ = n σ /σ 2 y 1 y −0. According to the Von Mises yield criterion. For a range of nratios. This function is the socalled plastic constraint factor pcf. plane strain (right) [39.5 σ /σ 1 1 1. 2 1.5 0 0. as expected. 9.7 Plastic constraint factor When the maximum principal stress is indicated as σmax = σ1 .5 1 0.89 Fig. we can consider the inﬂuence of the pcf on yielding.6 : Von Mises plane stress yield contour . 25. the other two pricipal stresses can be written as σ2 = nσ1 and σ3 = mσ1 . which is the ratio of the maximum principal stress at yield and the yield stress. the pcf is calulated and the corresponding {σ1 . σ2 }values constitute the plane stress Von Mises yield contour. 30]. σmax at yield can be expressed in the initial yield stress σy and a function of the ratios n and m.
such that the total yforce again equals the force associated with the elastic solution (= shaded area in the ﬁgure). m.1 Irwin plastic zone correction In 1958 Irwin [35] has devised a solution to this inconsistency. . 9.8. can be calculated. Irwin [36] suggests an over the thickness averaged value pcf = 2 2 = 1. The total internal force in ydirection is decreased. but experiments indicate that it is √ smaller. The ratio of σ3 and σ1 . 9. For ν = 3 its value is 3.90 In the crack tip region the principal stresses are known as a function of the angle θ around and the distance r to the crack tip. plane stress plane strain n = 1 − sin( θ ) / 1 + sin( θ ) 2 2 n = 1 − sin( θ ) / 1 + sin( θ ) 2 2 .68. plane stress pcf = 1 . is zero for plane stress and a function of Poisson’s ratio ν for plane strain. results in loss of equilibrium. based on the enlargement of the plastic zone. plane strain pcf = √ 1 1 − 4ν + 4ν 2 9. For plane stress we already saw that m = 0 and for plane strain m = 2ν. while the external force remains the same. it becomes clear that the requirement of σyy to be not larger than the yield stress σy . σxx σyy σy σxx σyy σy a ry r a ry rp r Fig. The plastic constraint factor can then be determined as 1 a function of Poisson’s ratio ν.7 : Irwin correction of plastic zone in the crack plane. m=0 θ m = 2ν/ 1 + sin( 2 ) In the plane of the crack where θ = 0 we have n = 1. The ratio of σ2 and σ1 .8 Plastic zone in the crack plane Considering the plastic zone in the crack plane (θ = 0) for plane stress state. n. .
load σ load σy KI (σ) = σ π(a + rp ) a + rp arccos π a a + rp KI (σy ) = 2σy . σ y σy x a rp σ Fig.8 : DugdaleBarenblatt correction of plastic zone in the crack plane. The elastic stresses for x > a + rp are known to be a series of terms of which the ﬁrst term is singular. Two load cases are considered to act on the crack with half length a + rp . applied to the crack faces over the distance rp and closing the crack to its initial half length a. where rp is the plastic zone length to be determined.91 θ=0 yield ry → KI σxx = σyy = √ 2πr → ry 0 σxx = σyy = σy KI σyy (r) dr = √ 2π ry = 1 2π KI σy → 2 2 σy rp = 0 1 2KI √ r − 2 dr = √ ry 2π √ 2KI ry rp = √ 2π σy → rp = 1 π KI σy = 2 ry 9. so leading to inﬁnite stress values. The ﬁrst load is the Mode I load σ. based on the requirement that in the elastic region outside the plastic zone the stresses must be ﬁnite. The second load is the yield stress σy .2 DugdaleBarenblatt plastic zone correction In 1960 Dugdale and Barenblatt have published a plastic zone calculation [15]. The requirement that this ﬁrst term is zero leads to the value of rp .8. 9. This ﬁrst term is completely determined by the addition of the two stress intensity factors of the two load cases.
criterion Von Mises Von Mises Tresca Tresca Tresca Irwin Irwin Dugdale Dugdale state plane stress plane strain plane stress plane strain σ1 > σ2 > σ3 plane strain σ1 > σ3 > σ2 plane stress plane strain (pcf = 3) plane stress plane strain (pcf = 3) 1 π 1 π π 8 π 8 ry or rp 1 2π 1 18π 1 2π 1 18π 0 KI σy KI 3σy KI σy KI 3σy 2 ry rp (KI /σy )2 2 KI σy KI σy KI σy KI σy 0.1592 2 0.0177 0 0.3927 2 0.0436 9. the boundary of the elasticplastic transition is considered for θ = 0. The table lists the distance to the crack tip of 3 the elasticplastic transition. as it is associated with a purely linear elastic solution.8.9 Small Scale Yielding The occurrence of a plastic zone at the crack tip rises the question whether the stress intensity factor K can still be used in a crack growth criterion. This situation is referred to as Small Scale Yielding (SSY). so in the plane of the crack.0354 2 0.0177 2 0.3 Plastic zones in the crack plane For comparison. can be used when the Kdomain is much larger than the plastic zone. . Although Irwin and Dugdale plastic zones were calculated for plane stress.3183 2 0.92 singular term = 0 → a = cos a + rp πσ 2σy KI (σ) = KI (σy ) → rp = 2 πKI 2 8σy → 9. The answer is that the elastic stress solution and thus its amplitude K. the plane strain equivalent results from the application of the plastic constraint factor as a magniﬁcation factor for the yield stress σy . Poisson’s ratio is ν = 1 .1592 2 0.
93 The accuracy of the elastic stress ﬁeld in SSY can be enhanced by using an eﬀective crack length. which takes the plastic zone into account. Calculation of K has to be done iteratively. aef f = a + (ry rp ) ↔ √ KI = βI (aef f )σ πaef f . and is calculated by adding ry or rp to a. as the plastic zone size depends on K and vice versa.
94 .
Chapter 10 Nonlinear Fracture Mechanics When Small Scale Yielding cannot be assumed or when the material behavior is intrinsically nonlinear. when plasticity at the crack tip is taken into account and the actual crack length is replaced by the eﬀective crack length. δt .1 Cracktip opening displacement In LEFM the displacement of material points in the region around the crack tip can be calculated. by taking r = a − x. where a is the half crack length. The origin of this xycoordinate system is at the crack center. the displacement uy in ydirection is known as a function of r (distance) and θ (angle). 10. r = a − x (1 + ν)(κ + 1) σ E 2 2a(a − x) uy = Crack Opening Displacement (COD) δ(x) = 2uy (x) = (1 + ν)(κ + 1) σ E 2a(a − x) Crack Tip Opening Displacement (CTOD) δt = δ(x = a) = 0 Wells (1963) [68] suggested that this CTOD can be used in a crack growth criterion. The displacement of points at the upper crack surface results for θ = π and can be expressed in the coordinate x. The crack opening (displacement) (COD) δ is two times this displacement. With the crack along the xaxis. both for plane stress and for plane strain. is zero. 95 . concepts from LEFM loose their meaning. √ r σ πa sin( 1 θ) κ + 1 − 2 cos2 ( 1 θ) uy = 2 2 2µ 2π displacement in crack plane θ = π. For such cases NonLinear Fracture Mechanics (NLFM) or ElastoPlastic Fracture Mechanics provides new theories and concepts to analye the behavior of cracks. It is obvious that the opening at the crack tip (CTOD).
10.1.96 10. which is the actual crack length plus the length of the plastic zone in front of the crack tip. Be aware that ry is used here and not the corrected plastic zone rp . (1 + ν)(κ + 1) σ E (1 + ν)(κ + 1) σ = E δ(x) = 2aef f (aef f − x) 2(a + ry )(a + ry − x) δt = δ(x = a) = (1 + ν)(κ + 1) σ 2(a + ry )ry E (1 + ν)(κ + 1) 2 σ 2ary + 2ry = E (1 + ν)(κ + 1) ≈ σ 2ary E 2 4G 4 KI = π Eσy π σy 2 4(1 − ν 2 ) KI 1 δt = √ π Eσy 3 plane stress plane strain : : δt = .1 : Irwin plastic zone correction. σxx σyy σy σxx σyy σy a ry r a ry rp r Fig.1 CTOD by Irwin The inﬂuence of the crack tip plastic zone can be taken into account by using an eﬀective crack length aef f . 1 2π KI σy 2 aef f = a + ry = a + When the eﬀective crack length of Irwin is used. the crack tip opening can be calculated for plane stress and for plane strain. The correction of Irwin is illustrated in the ﬁgure.
displacement from requirement ”singular term = 0” uy (x) = ¯ (a + rp )σy πE x ln a + rp x a + rp .2 : Dugdale plastic zone correction. results in a crack tip opening which is not zero. the length of the plastic zone follows from the requirement that the stress at the elasticplastic boundary is not singular.97 10. 10.2 CTOD by Dugdale In the vision of BarenblattDugdale.1. aef f π = a + rp = a + 8 KI σy 2 The requirement that the singular term in the stresses at the crack tip must be zero. It can be calculated for plane stress and for plane strain. σ y σy x a rp σ Fig. sin2 (ˆ − γ) γ 2 sin (ˆ + γ) γ γ= ˆ π σ 2 σy : uy (x) ¯ sin(ˆ ) + sin(γ) γ sin(ˆ ) − sin(γ) γ 2 + cos(ˆ ) ln γ γ = arccos Crack Tip Opening Displacement 8σv a δt = lim 2¯y (x) = u ln sec x→a πE series expansion & σ ≪ σy 2 KI G = Eσy σy K2 1 (1 − ν 2 ) I δt = 2 Eσy π σ 2 σy → plane stress plane strain : : δt = .
with three components in a threedimensional (here Cartesian) coordinate system.2 Jintegral In LEFM. which has been introduced by Rice in fracture mechanics in 1968 [54]. J . W = 0 σij dεij N m J = J1 = Γ W n 1 − ti . which can be compared to a critical value in a crack growth criterion. n t x2 Ω V e2 Γ positive e1 x1 S Fig. has to be measured.3 : Closed curve in a plane with unit normal vector and stress vector. Also the unit normal vector n in that point. the speciﬁc elastic energy must be calculated from stresses and strains.98 10. the behavior of a crack can be described with the Jintegral. 10. What is generally called the Jintegral in fracture mechanics. In NLFM the CTOD is a measure for the deformation at the crack tip. In each point of this trajectory. Its value results from an integration along a trajectory Γ . In NLFM. δt = δtc (ε. crack growth can be predicted after calculation of the energy release rate G or the stress intensity factor K. the stress vector t and the spatial derivatives of the displacement u play a role.3 CTOD crack growth criterion In LEFM the CTOD can be related to the energy release rate G and the stress intensity factor KI . T ) ˙ 10. is the ﬁrst component of the vector J. The critical value. which may depend on strain rate and/or temperature. so the component in x1 direction.1. Epq Jk = Γ W n k − ti ∂ui ∂xk dΓ ∂ui ∂x1 dΓ . The Jintegral is a vector.
kj dΩ = 0 dΩ = σmn um. This important result means that the integral over a nonclosed path around the crack tip. which runs around the crack tip and along both crack surfaces.2.k nj dΓ Stokes (Gauss in 3D) dΩ inside Γ no singularities → Ω dW ∂εmn δjk − σij. When the movement over the line is anticlockwise.j = 0 + un.j ui. On the crack faces. when it is clockwise.k − σij ui. and the Jintegral is the sum of two terms representing an integration along two lines around the crack tip. This result of the integration along the crack faces is zero. . When this path is closed. It is also necessary to use a linear straindisplacement deﬁnition. it can be proved that the resulting value is always zero. The Jintegral – ﬁrst component of the vector – is calculated along a closed path.kj Ω 10.1 Integral along closed curve The Jintegral vector is deﬁned by integration along a path in a twodimensional plane. always has the same value and so the Jintegral is path independent.kj dεmn ∂xj σmn = ∂W ∂εmn homogeneous hyperelastic linear strain equilibrium equations 1 2 σmn (um. The integral can be written as a summation of four parts. It is assumed that the crack faces are not loaded (ti = 0i ).nk − σij ui.nk Ω 1 εmn = 2 (um.m ) σij.2. it appears that the value of both integrals has to be equal. With this sign convention. the material is not subjected to volume loads and the acceleration is zero. The prove assumes that the material behavior is hyperelastic and homogeneous. Jk = Γ W δjk − σij ui. we assign the value a positive value and.n + un.mk ) − σij ui. Obviously the calculated value is zero. which is only allowed when deformation and rotations are small. starting at one crack face and ending at the other.99 10.2 Path independence A crack is now introduced along the x1 axis. the unit normal vector is pointing in x2 direction. so its ﬁrst component n1 is zero. under the conditions that there are no singularities in the area within the closed path. the result is negative.
2. Γ+ no loading of crack faces : ti = 0 on f1 dΓ = 0 ΓB ΓA ΓA f1 dΓ = −J1A . lin.4 : Closed curve including crack surfaces. material : Jk = Γ W = 1 σmn εmn = 1 σmn (um.k nj dΓ = Γ 1 2 σmn um.n δjk . they are characterized by the stress intensity factors KI .m ) 2 4 1 4 σmn (um.n + un. elast.m )δjk − σij ui. the stress and displacement components at the crack tip are known as a function of the position relative to the crack tip. the integration path can be chosen to be a circle with the crack tip as its center. For multimode loading. Integration over this circular path reveals that the Jintegral is related to the stress intensity factors. KII and KIII .n + un. 10.3 Relation J ∼ K In LEFM. fk dΓ + ΓA ΓB fk dΓ + Γ− fk dΓ + f1 dΓ + Γ− Γ+ fk dΓ = 0 f1 dΓ = 0 Γ + and Γ − → f1 dΓ + ΓA ΓB f1 dΓ + n1 = 0 f1 dΓ + .100 x2 Ω e2 e1 Γ+ Γ− n x1 ΓB ΓA n Fig. Because the Jintegral is pathindependent.k nj dΓ − σij ui. ΓB f1 dΓ = J1B −J1A + J1B = 0 → J1A = J1B 10.
111 + 0.101 Mode I + II + III 1 σij = √ [KI fIij + KII fIIij + KIII fIIIij ] 2πr ui = uIi + uIIi + uIIIi substitution and integration over Γ = circle (κ + 1)(1 + ν) (1 + ν) 2 2 2 KIII KI + KII + 4E E (κ + 1)(1 + ν) KI KII J2 = − 2E J1 = For Mode I loading of the crack. a better relation is suggested by ASTM. This means that the Jintegral can be used in the crack growth criteria of LEFM as a replacement for K and G. when the material behavior is described by the RambergOsgood relation [31. Jintegral is now also related to the crack tip opening displacement δt . it follows immediately that the Jintegral is equivalent to the energy release rate G.3 HRR crack tip stresses and strains Hutchinson. . Rice and Rosengren (HRR) derived a solution for the crack tip stress and displacement components. J = m σy δt m = −0. 55]. Mode I plane stress plane strain (κ + 1)(1 + ν) 2 KI = G 4E 3−ν 1+ν 4 1 2 κ+1 = + = → J= K 1+ν 1+ν 1+ν E I (1 − ν 2 ) 2 KI κ + 1 = 4 − 4ν → J= E J1 = J = Obviously. plane stress Irwin Dugbale Irwin Dugbale π J = σy δt 4 J = σy δt π√ J= 3 σy δt 4 J = 2σy δt plane strain Based on emperical observations.817 a σu + 1. including the inﬂuence of the specimen dimensions and the ultimate strength σu .36 W σy 10.
For n > 1 the singularity is smaller. Although it is a general nonlinear material model. ui = αεy0 β n r n+1 ui (θ) ˜ 1 n+1 J β= ασy0 εy0 In (In from num. σ ε = +α εy0 σy0 5 n=1 σ σy0 n Ramberg−Osgood for α = 0. the same as we had in LEFM. describing the linear elastic behavior. The model has two parameters.102 10. The solution reveals that the dependency of the distance to the crack tip is determined √ by the exponent n.3.01 4 n=3 3 σ/σy0 n=5 2 n=7 n = 13 1 0 0 1 2 3 ε/εy0 4 5 6 Fig. both for plane stress and plane strain. just as it is by K in LEFM. The parameter β also depends on a constant In . For n = 1 the singularity of the stresses is 1/ r. The initial yield stress is σy0 and the initial yield strain is εy0 = σy0 /E.01. a coeﬃcient α and an exponent n.) . 10.1 RambergOsgood material law Nonlinear material behavior can often be described by an exponential relation between (equivalent) stress and (equivalent) strain. A much used relation of this type is the RambergOsgood relation. where E is Youngs modulus.3. which value is determined numerically as a function of the exponent n. When n = 1 the relation describes linear hardening and with n→∞ we have ideal plastic behavior as is illustrated in the ﬁgure for α = 0. 10. which is the strain hardening parameter. we will use it here to describe postyield behavior.5 : Stressstrain relation according to the RambergOsgood material law. anal. This means that the stress and deformation state at the crack tip is completely characterized by J. which is directly related to the Jintegral. 1 − n+1 1 σij = σy0 β r with : σij (θ) ˜ .2 HRRsolution The HRRsolution comprises a parameter β.
103 In
5 2.5 0 5 10 15
plane strain plane stress
n
Fig. 10.6 : Value of the variable In for plane strain and plane stress.
10.3.3
Jintegral crack growth criterion
We saw that the Jintegral can replace the energy release rate in LEFM and is related to the stress intensity factor. In NLFM, where the material behavior is described by the general RambergOsgood relation, the Jintegral characterizes the stress at the crack tip. It is thus obvious that it can be used in a crack growth criterion. Calculation of its value is easily done, due to the fact that the integration path can be chosen arbitrarily. Critical values have to be measured according to normalized experiments.
J = Jc
104
Chapter 11
Numerical fracture mechanics
In the crack growth criteria of LEFM and NLFM, a crack growth parameter has to be calculated and compared to a critical value, which is known from experiments. Crack parameters, G, K, δt and J, can be calculated analytically, but this is only possible with great mathematical eﬀort and even then for rather simple cases. When such solutions exist, they can often be found in literature. Using numerical methods, it is possible to calculate their values rather straightforwardly, also for practical problems. Finite element programs are widely available and can be used for the numerical calculation of crack growth parameters. Although special procedures are developed and implemented, the use of standard elements is possible after a minor adaptation of the element mesh. Calculation of the Jintegral is often implemented as a standard option. With the ﬁnite element method it is also possible to simulate crack propagation. Several techniques are used and new ones are still under development. Besides the ﬁnite element method, the boundary element method is used in numerical fracture mechanics. Commercial programs for BEM are available although not so widespread as for FEM.
11.1
Quadratic elements
For twodimensional ﬁnite element analyses, the 8node plane stress or plane strain element gives accurate results for most mechanical problems. The displacement of an internal element point is interpolated quadratically between the nodal displacements. Although initially the element edges are straight lines, after deformation they may become of parabolic shape. Interpolation or shape functions are used for this purpose, which are a function of two local socalled isoparametric coordinates ξ1 and ξ2 , which have values between 1 and +1. The ξ1 , ξ2 coordinate system is not orthonormal, because the axes are not perpendicular. Strains, which are derivatives of the displacements, are more or less linear in the local coordinates and this is also the case for the stresses, which are proportional to the strains. It has to be said that these statements are not valid when nonlinear straindisplacement and stressstrain relations are used.
105
which is not described by the element interpolation functions.e. Fig. 11. it is found that reﬁning the mesh at the crack tip will not lead to convergence. However. 11. further reﬁnement will always render diﬀerent results. 11. The very steep stress gradient at the crack tip can only be described with a lot of small elements.106 ξ2 4 8 1 3 7 6 6 5 2 1 8 5 2 5 3 7 4 7 ξ2 ξ1 3 ξ1 6 2 4 8 1 Fig.1 : Quadratic elements with 8 nodal points. The reason for this unacceptable mesh dependency is the singularity in the stress ﬁeld. theory says that the stresses at √ the crack tip are proportional to 1/ r.2 : Crack tip mesh. . Where the crack has opened. with r the distance to the crack tip. the nodes of elements above and below the crack may coincide. but must not be connected.2 Crack tip mesh The ﬁgure shows part of the element mesh around the crack tip. i. When the material is assumed to be linearly elastic.
However. where two midside points are repositioned towards one corner node. 4 7 3 3 6 5 6 4 2 4 8 1 7 p 5 3p 3 6 2 4 1 2 3 8 1 5 p 3p 2 1 Fig. They are standard 8node quadrilateral or 6node triangular elements. It can be proved that they √ describe the 1/ rstress ﬁeld singularity at the crack tip accurately. This results in a socalled enriched element. The big advantage is that they allow the use of standard ﬁnite element packages. which is only usable at the crack tip. One solution is to change the interpolation function of the element in such a way that the crack tip ﬁeld is included. Quarter point elements are also referred to as Distorted Quadratic Quadrilateral or Triangle. 11. Another solution may be found in application of a modiﬁed variational principle as the basis for the ﬁnite element equations. Using enriched or hybrid elements implies implementation of these elements and sometimes new solution techniques in the FEM program. which may be very unfeasable. which is located in the node towards which the midside nodes are moved. where three points of the 8node quadrilateral element are situated at the same location. Quarter point elements can be made very easily at the crack tip by replacing some midside nodes of the element mesh. which are slightly adapted into socalled quarter point elements. there is a possibility to calculate the crack tip stress ﬁeld very accurately with standard 8node elements. when a commercial program is used. but not connected.107 11. such that they divide there side in the ratio 1:3. The resulting element is a hybrid element. . A linear 4node quadrilateral element can be collapsed into a triangle and it appears to describe a 1/r singularity very accurately.4 Quarter point elements Quarter point elements are devised by Barsoum in 1976 [2] as a simple means to describe the stress singularity at the crack tip. They are used at the crack tip. are still being sought and found. 11. where stresses are unknown nodal variables. Another suitable solution is to use a Collapsed Quadratic Quadrilateral.3 Special elements Solutions to circumvent the shortcomings of the standard 8node element to describe the crack tip stresses well.3 : Quarter point and collapsed elements.
11. using the interpolations. to use transition elements between the quarter point elements and the regular mid point elements for better results. The ﬁgure shows a line element with three nodes and a local coordinate −1 ≤ ξ ≤ 1. which can be calculated.4 : Crack tip rozet. The displacement u in xdirection is interpolated in the same way between the nodal displacements u1 . 11.1 Onedimensional case For a onedimensional case. the length of the element. The global position of an element point can be interpolated between the position of the three nodes {x1 . x3 }. Fig. it will be proved that the quarter point technique indeed renders √ the 1/ r singularity.5 : Onedimensional threepoint element at the crack tip. 11.r. x. Lim advised in 1993 [43].t. x2 . u2 and u3 . ξ=0 3 ξ=1 2 1 1 x = 2 ξ(ξ − 1)x1 + 2 ξ(ξ + 1)x2 − (ξ 2 − 1)x3 1 = 2 ξ(ξ + 1)L − (ξ 2 − 1)x3 . The strain of the element is calculated as the derivative of u w.4. The global coordinate x has its origin in node 1.108 The crack tip mesh is often made as a crack tip rozet. where the crack tip can be seen. ξ = −1 1 x Fig. where x1 = 0 and x2 = L. with quarter point elements at the crack tip.
1 1 x = 2 ξ(ξ + 1)L − (ξ 2 − 1) 1 L = 2 (ξ + 1)L 2 ⇒ 1 2 dx 1 = 2L dξ u2 + 2u3 du dξ = 1 dx 2L du → du dx ξ=−1 x=0 = 2 L 3 − 2 u1 + Quarter point element 1 When node 3 is positioned at the quarter of the element length (x3 = 4 L). the strain appears √ to be proportional to 1/ x and becomes inﬁnite in node 1.109 1 1 u = 2 ξ(ξ − 1)u1 + 2 ξ(ξ + 1)u2 − (ξ 2 − 1)u3 du du dξ du dx du 1 → = (ξ − 2 )u1 + (ξ + 1 )u2 − 2ξu3 = = / 2 dξ dx dξ dx dξ dξ Mid point element When node 3 is located in the center of the element (x3 = 1 L). ξ = −1 1 ξ=0 3 x ξ=1 2 Fig. ξ = −1 1 x ξ=0 3 ξ=1 2 Fig. 11. The strain at node 1 (the crack tip) is constant. 11. 1 1 x = 1 ξ(ξ + 1)L − (ξ 2 − 1) 4 L = 4 (ξ + 1)2 L 2 √ dx = 1 (ξ + 1)L = xL 2 dξ du dξ =√ dx xL du → ξ+1= 4x L ⇒ → du dx ξ=−1 x=0 =∞ → singularity 1 √ x .6 : Onedimensional mid point element at the crack tip. the strain in node 1 (at the 2 crack tip) can be evaluated and appears to be a linear function of the local coordinate ξ.7 : Onedimensional quarter point element at the crack tip.
. it is important to use quarter point elements at the crack tip. where the outer boundary of the crack tip mesh is ﬁxed. The ratio of the diﬀerence in elastic energy and ∆a is a good approximation for G.r. when ∆a is indeed very small and when ”ﬁxed grips” is used. This procedure is called the Virtual Extension Method or Stiﬀness Derivation Method. 11. Because only a small region around the crack tip is considered. The change in the stiﬀness matrix can be used directly to evaluate G. It is deﬁned by diﬀerentiation of internal and external work w. Only a small part of the mesh around the crack tip is considered and its contribution to the global stiﬀness matrix C is calculated for crack length a and for crack length a + ∆a.5 Virtual crack extension method (VCEM) One of the variables which can be calculated with a standard ﬁnite element package is the energy release rate G. with a small diﬀerence in crack length ∆a.e. u u a a + ∆a I II Fig. Fig. i.9 : VCEM in crack tip mesh. when boundary displacements are prescribed.t. the crack length. The energy release rate can be calculated following two analyses.110 11. ﬁxed grips → dUe =0 da ⇒ G=− 1 dUi 1 Ui (a + ∆a) − Ui (a) ≈− B da B ∆a stiﬀness matrix variation Instead of analyzing the whole structure two times with small diﬀerence in crack length. 11. the energy release rate can be calculated directly from the stiﬀness matrix.8 : Virtual Crack Extension Method.
. 11. √ √ 2πr σ12 θ=0 KI = lim r →0 2πr σ22 θ=0 . KII = lim r →0 p1 p2 p3 p4 θ r K Kp1 Kp2 Kp3 Kp4 r r1 r2 r3 r4 Fig. plane stress plane strain 2 KI = EGI E 2 KI = GI 1 − ν2 . 2 KII = EGII E 2 KII = GII 1 − ν2 Calculation of the stress intensity factor is also possible through its deﬁnition. 4(1 − ν 2 ) E uy = r KI gij (θ) 2π → E KI = lim r →0 4(1 − ν 2 ) 2π uy (θ = 0) r . Stresses in element integration points are calculated by diﬀerentiation.10 : Calculation of K by extrapolation. Because stresses are calculated in integration points and the crack tip is located in an element node. this calculation is not exactly according to the deﬁnition of K. It always remains to be investigated how the interpolation has to be done to get the best results.6 Stress intensity factor Calculation of the stress intensity factor is of course possible trough its relation with the energy release rate. which relates it to the stress ﬁeld. which makes them less accurate then the calculated nodal displacements. Stress components have to be calculated accurately upon which K results from extrapolation to the crack tip. It seems logical therefore to calculate the stress intensity factor not from the stress ﬁeld but from the displacement ﬁeld.111 BG=− dUi ∆C = − 1 uT u 2 da ˜ ∆a ˜ with ∆C = C(a + ∆a) − C(a) 11.
It consists of lines in coordinate directions. published in 1985. When such a path is taken at a certain distance from the crack tip. it is shown that the transformation is possible when a shift function q is introduced. 11. In a paper by Li et.112 11. 11. by integration along a path around the crack tip. Because the path Γ is not a closed path.12 : Jintegral path through integration points. the integration path must preferably pass through these points. In a general crack tip mesh such a path would be diﬃcult to deﬁne during postprocessing of the results.11 : Jintegral path in crack tip rozet.7. only half of the path has to be taken into account. i. enclosed by Γ . there is no need to use quarter point crack tip elements. In a regular crack tip mesh. . Fig. 11.1 Domain integration In commercial element packages. Stokes theorem cannot be applied directly. an integration path could be deﬁned as indicated in the ﬁgure. Fig. which is 0 at the outer boundary of Ω and 1 at the inner boundary.al. the Jintegral over a curve Γ is transformed to an integral over the domain Ω. Because stresses and strains are known in integration points. When symmetry can be considered.7 Jintegral The Jintegral can be calculated according to its deﬁnition.e.
which includes phenomena like pressure on the crack faces pi . Fig. In one analysis several options can be chosen. Calculating such Jintegral values is a default option in major FEA codes.7. 11. J= Ω ∂q ∂xj σij ∂ui − W δ1j ∂x1 dΩ 11.14 : Jintegral domain with various element rings. The user has to indicate which ring of elements around the crack tip must be moved when the crack is virtually extended. It is calculated as ij a domain integral. 11.2 De Lorenzi Jintegral : VCE technique In the Marc/Mentat package a general Jintegral is calculated. It does not require intensive mesh reﬁnement near the crack tip or the use of quarter point elements. where the shift function q is deﬁned as the ratio of a displacement ∆x1 of nodes within the domain Ω and the elongation ∆a of the crack. acceleration eﬀects and initial strains εo .13 : Jintegral domain.113 x2 Ω e2 e1 Γ+ Γ− n ΓB x1 ΓA Ω n q=0 q=1 Fig. .
The actual crack propagation can be simulated.114 J= Ω ∂q ∂xj qpi Γs σi1 ∂ui − W δ1j ∂x1 dΩ − ∂ui dΩ + ∂x1 qσij Ω ∂ui dΓ − ∂x1 q(ρqi − ρ¨i ) u Ω ∂εo ij dΩ ∂x1 11. 11. reveals in what direction the crack tip wants to move. small elements are required there. 11.8. By introducing springs and/or dashpots between the released nodes. 11. which are not connected anymore. Fig.15 : Node release method for crack growth simulation. Various approaches exist and they are improved in ongoing research. When crack propagation is simulated.2 Moving Crack Tip Mesh Because of the high stress gradients at the crack tip. In the node release approach.8 Crack growth simulation Numerical procedures can not only be used to evaluate crack growth criteria. . The resulting crack path is strongly depending on the mesh geometry. a crack can only move along element edges and the movement is realized by splitting one or more nodes in two.8. while elements at a some distance to the crack tip can be considerably larger.1 Node release One of the techniques to simulate crack propagation is the ”node release” method. use of a crack growth direction criterion like MTS or SED. ductile crack propagation can be simulated. It is for this reason that a socalled crack tip mesh (CTM) is often used. because the propagation implies that the element mesh topology has to change. When a crack growth criterion indicates that a crack will growth. Application is not so easy to accomplish.
which initially was in the ﬁbermatrix interface but from some point moved into the matrix material. The CTM was moved by hand for each crack advancement. The example shows the crack propagation in the interface between ﬁber and matrix for a plane strain simulation in the transversal plane. 11. which changes the mesh topology. It was possible to simulate the crack propagation.115 this CTM can follow the crack tip. . Mesh reﬁnement can be realized in the region of the crack tip. Fig. The crack opening is shown with enlarged displacements.16 : Crack tip mesh moves along ﬁbermatrix interface [42]. It implies the subdivision of an element in two or more smaller elements.8.3 Element splitting Instead of advancing a crack along element edges. Results are more accurate than those of the node release method. which was a very laborious operation. propagation through elements is also an option. 11. where the loading is in the vertical direction.
but the material damage is modeled with small cracks in element integration points. These cracks may initiate and lengthen according to certain loading criteria.17 : Crack propagation by element splitting technique [46]. 11. but simulated with anisotropic material behavior.8.4 Smeared crack approach The smeared crack approach can be seen as an intermediate between fracture mechanics and continuum damage mechanics.116 Fig. where in one direction the material stiﬀness is reduced. 11. Attention is not focused on one large crack. . The cracks are not really modeled.
18 : Smeared crack approach for crack propagation. .117 e2 e1 e2 n2 σ2 e1 n2 n1 σ1 n1 Fig. 11.
118 .
1 Crack surface When fatigue failure has occurred. Fatigue failure became an issue when carriages and trains with iron or steel axles began traveling frequently over long distances in the mid 19th century. all of course o experimentally. structural parts fail unexpectedly. 119 . a tremendous amount of data was gathered. random forces (ships. socalled clam shell markings also called beach marks can be seen. W¨hler was among the ﬁrst to investigate these phenomena systematically. Over many years. often leading to dramatic and overall collapse. but theoretical explanation had to wait until Griﬃth saw the importance of cracks for the failure of materials. due to changes in loading conditions. 12. vehicles. thermal cycling (switching of electronic devices). with an amplitude so low that failure due to yielding or crack growth. is never to be expected. planes) occur always and everywhere. At a macroscopic scale. Cyclic loading of structures does not only occur as the result of deliberately applied load reversals. the crack displacement during each individual load cycle is seen as a pattern of striations. experience tells a diﬀerent story. Axles broke mostly at diameterjumps and this was explained ﬁrst as the result of changes in the material’s structure. During one cycle. the crack surface shows two distinct markings on two diﬀerent scales. This very dangerous phenomenon is the result of fatigue crack growth. When the crack surface is observed with a microscope.Chapter 12 Fatigue Many structures are subjected to cyclic loading. plastic deformation at the crack tip results in sequential blunting and sharpening of the crack tip. repeated pressurization en depressurization (airplanes). After a large number of load reversals. Vibrations (machine(s) (parts)). However. They are the result of irregularities in the growth of the fatigue crack. This crack growth is very regular.
because size eﬀects and upscaling of load amplitude and frequencies may lead to inconsistencies with reallife applications. is often not feasible. due to costs and the extremely long times involved before the detection of fatigue damage. all these variables obviously apply to strains. For stress controlled loading. 12. Experiments are therefore mostly done on laboratory scale. Results must be carefully evaluated. .1 : Clam shell fatigue crack surface and striations [29]. using small test specimens subjected to high loads varying at high frequencies. Also the amplitude ratio σa /σm is sometimes used.2 : Harmonic stress. σ σmax σm σmin 0 0 i i+1 t N Fig. In strain controlled experiments. the loading is ﬁrst taken to be harmonic and is characterized in terms of stresses or strains.3 Fatigue load When fatigue crack growth is studied experimentally. 12. 12. The stress amplitude σa is half the stress variation.2 Experiments Experiments can be done on fullscale structures subjected to reallife loading. however. the stress range ∆σ is the diﬀerence between the maximum and the minimum stress during the cycle. 12. This. Average stress σm and stress ratio Rσ are used to specify the loading.120 Fig.
some Ti/Al/Mg alloys – do not have such a fatigue limit: they will always fail no matter how low the stress. 3518 part I 1984. When a fatigue limit does not exist. some Ti/Al/Mg alloys – never fail under cyclic loading.3 : Harmonic stress and fatigue limit. some austenitic steels.121 ∆σ = σmax − σmin σm = 1 (σmax + σmin ) 2 . σ σth N Fig.g.3. low strength alloys. Other materials – high strength steels. . Determination and representation of these data is prescribed in normalization sheet B.S. the (SN)curve will advance towards a horizontal asymptote at the level σ = σth . . when the stress in each cycle is below a certain threshold σth . σa = 1 ∆σ 2 Rσ = σmin /σmax . 12. 1−R σa = σm 1+R 12. These curves are mostly recorded for zero average stress. mild steel.1 Fatigue limit Experiments reveal that some materials – e.or W¨hler curve. For materials with a fatigue limit. The ratio of fatigue limit or endurance limit to tensile strength is referred to as endurance ratio.3. 12. where the maximum o stress S = σmax is indicated along the vertical axis and the logarithm of the number of cycles to failure or fatigue life Nf along the horizontal axis. although it may take a very large number of load cycles.2 (SN)curve The experimental data are represented as a (SN). also referred to as fatigue limit. most nonferro alloys. the fatigue strength or endurance limit is deﬁned as the stress variation for failure after a speciﬁed high – typically 106 – number of cycles.
12.3. indicating that the fatigue life Nf is reached earlier for the same stress ∗ amplitude σa . 12. σa is reduced to σa . It can also be stated that for the same Nf . 600 550 500 450 400 σmax [MPa] 350 300 250 steel1020 200 150 Mgalloy 100 4 5 6 7 8 9 steelT1 Al2024T4 10 10 10 10 N f 10 10 Fig.4 : SNcurve with fatigue limit.3 Inﬂuence of average stress When the average stress is positive.122 S σth 0 0 log(Nf ) Fig. a crack will grow faster and the curve will be located at lower stress values. Fatigue data may also be presented in a (Sa N)curve. . where the vertical axis indicates the stress amplitude : Sa = σa = 1 ∆σ. 2 For some real materials the (SN)curves are shown in the ﬁgure.5 : SNcurves for various materials. 12.
7 : PSNcurves. Corrections have been proposed to take into account the decrease in σa due to σm = 0 and to relate it to the tensile strength σu or the initial yield stress σy0 . The center curve represents the stress for which 50 % of all specimens will fail after the associated Nf . Much used are the emperical formulas of Gerber. 550 500 450 400 350 σmax [MPa] 300 90% prob. 12.6 : Inﬂuence of the average stress on the Sa Ncurve.123 σa σm σth 0 0 log(Nf ) Fig. 12.4 (PSN)curve Variations in experimental fatigue data can be represented in a (PSN)curve.failure 250 200 150 10% prob.3. Goodman and Soderberg. 1 % of all specimens fail for Nf on this curve. Gerber (1874) Goodman (1899) Soderberg (1939) ∗ σm σa =1− σa σu ∗ σa σm =1− σa σu ∗ σm σa =1− σa σy0 2 12. . where probability (P) of failure is incorporated. The lower curve represents the weakest specimens. When the upper curve is reached.failure 100 10 4 10 5 10 N 6 10 7 10 8 f Fig. 99 % of all specimens will have failed.failure 50% prob. For the given stress.
and the fatigue strength exponent or Basquin exponent b. For low cycle fatigue the MansonCoﬃn relation is used for the ﬁt. the plastic crack tip zone is small and concepts from LEFM (+ SSY) can be applied. For high cycle fatigue the (SN)curve can be ﬁtted with the Basquin relation. we are in the High Cycle Fatigue (HCF) range.8 : Low and High Cycle Fatigue regimes.9 : Basquin ﬁt of SNcurve. . Crack tip stresses are so high that NLFM has to be applied to study the crack behavior. 12. 1 2 ∆σ ′ = σa = σf (2Nf )b → −b ∆σNf = constant log ∆σ 2 log(2Nf ) Fig.3. we are in the Low Cycle Fatigue (LCF) range. which is twice the number of cycles. For high stresses and thus short fatigue life. 12.124 12. Because stresses are very low.5 High/low cycle fatigue For low stresses and thus long fatigue life (typically Nf > 50000). 12.6 Basquin relation The Basquin relation relates the stress amplitude to the number of load reversals. We can characterize the crack tip stress ﬁeld with stress intensity factors. Sa σm = 0 0 0 4 LCF 5 HCF log(Nf ) Fig. The relation has two parameters: the fatigue strength coeﬃcient ′ σf .3.
8 Total strainlife curve Basquin and MansonCoﬃn relations can be combined. The relation has two parameters: the fatigue ductility coeﬃcient ε′ . which is easily accomplished. which is twice the number of cycles.10 : MansonCoﬃn ﬁt of SNcurve. which is a valid assumption for small deformations. and the fatigue ductility exponent or MansonCoﬃn exponent c.7 MansonCoﬃn relation The MansonCoﬃn relation relates the plastic strain amplitude to the number of load reversals.3.125 12. log( ∆ε ) 2 log(Nf ) Fig. because the low stresses assure linear elastic behavior. 12.3. using logarithmic axes. the stress amplitude in the Basquin relation has to be replaced by a strain amplitude. 12. 12. The total strain amplitude is the addition of the elastic and plastic strain amplitudes. The resulting total strainlife curve can be plotted.11 : Total strainlife curve. ∆ε ∆εe ∆εp 1 ′ = + = σf (2Nf )b + ε′ (2Nf )c f 2 2 2 E . To that purpose. f p 1 2 ∆ε = ε′ (2Nf )c f → −c ∆εp Nf = constant log ∆εp 2 log(2Nf ) Fig.
∆σth (notched) = 1 ∆σth (unnotched) Kf . almost always leads to inhomogeneous deformation and nonuniform stress ﬁelds. however. Manufacturing procedures are also very inﬂuential. The inﬂuence of the average stress σm is obvious and has been discussed.4.2 Stress concentrations The shape of structural parts. combined with the external loading. Shear loading experiments result in diﬀerent (SN)curves than the tensile equivalents. For real structures. In certain situations local stresses can become much higher than the average or nominal stress and such stress concentrations are mostly found at notches. Relations for q have been proposed by Peterson [53] and Neuber [50]. where Kf is the fatigue strength reduction factor. the fatigue or endurance limit is theoretically reduced with 1/Kt .4 Inﬂuence factors Fatigue life of a test specimen is determined by the external load amplitude and frequency. The notch sensitivity factor q relates Kf to Kt and is found to be a function of the notch radius and a material parameter. 12. as is experimentally represented in an (SN)curve. Environmental inﬂuences can have a high impact on the initiation and propagation of fatigue damage as well. the fatigue life is highly determined by macroscopic shape and dimension.1 Load spectrum Fatigue life is experimentally represented in an (SN)curve. that the fatigue life is reduced less severely than this: it is reduced by a factor 1/Kf . When the specimen is subjected to an harmonic load. Experiments have shown. Too low or too high frequencies will result in diﬀerent fatigue life. Fatigue data are independent of the loading frequency within certain limits. Multiaxial loading results in lower fatigue life that uniaxial loading. Until so far we considered tensile loading where S can be the maximum stress or the stress amplitude. 12. and by microscopic conﬁguration (material microsructure).12 : Stress concentration at axis diameter reduction. The stress concentration factor Kt is the ratio between the maximum stress and the nominal stress and is a function of the notch radius ρ. which is smaller than Kt .126 12. 12. ρ Fig. 1 < Kf < Kt .4.
3 Stress gradients The size of a structural part. so fatigue limit is lower for small grains. particles. The texture (grain shape) also inﬂuences fatigue life and leads to anisotropy. When the gradient is lower. an axis subjected to bending. This results in a decrease of fatigue life with specimen size.127 Kf = 1 + q(Kt − 1) Peterson : Neuber : q= q= 1 1+ a ρ 1 1+ b ρ with with a = material parameter b = grain size parameter 12.4. Inhomogeneities – lattice imperfection. At low temperatures grain boundaries act as a barrier for crack propagation. determines the gradients of stresses at equal external load. Fig. 12. more material is subjected to a raised stress level than in the case of a higher gradient. ﬁbers – may have a negative (stress concentration → initiation of cracks) or positive (crack arrest.4 Material properties The ease of fatigue crack initiation and growth is obviously inﬂuenced by the microstructure of the material. voids. e. the grain boundaries are weakened by diﬀusion of atoms (creep).4.g. when subjected to the same external load. 12.13 : Stress gradients in axes of diﬀerent diameter. especially for critical applications like aviation. dissipation) inﬂuence on fatigue life. This size eﬀect often necessitates testing in full scale experiments. Larger specimens show lower gradients than smaller ones. so fatigue limit is higher for small grain microstructures. . At high temperatures.
6 Environment Environmental inﬂuences on fatigue life are mostly due to temperature eﬀects and chemical attack. Hydrogen and other gasses can diﬀuse to the grain boundaries and diminish bond strength.and extrusion of lattices at the surface.4. crystal lattices may be squeezed out by cyclic plastic deformation. where stress concentrations will easily lead to the initiation of a fatigue crack. internal surfaces at bulk defects.4. will result in a higher resistance against lattice slip. Surface treatments. A coating can be applied to eliminate deteriorating environmental eﬀects. They show up at about 0.5Tm ) facilitates the creep deformation due to diﬀusion of atoms to and from grain boundaries. Fatigue cracks will travel more easily between grains or cross grain boundaries. This means that crack growth is very slow until the ﬁnal stage in the . High temperature (T > 0. The resulting extrusions (and intrusions) act as a notch. like shot peening and carbonizing. can be applied to induce compressive stresses. due to the prevention of plastic deformation. 10µm Fig.5 Surface quality Material microstructure determines the growth of a fatigue crack to a high extent. 12. dislocation movement and migration of vacancies.14 : In. Protective surface ﬁlms may be cracked. Grain boundaries may also be weakened by chemical inﬂuence. The surface roughness can be set during manufacturing (or specimen preparation) and will highly inﬂuence the surface crack initiation and fatigue life.1 Nf . Low temperatures lead to embrittlement.5 Crack growth Detailed experiments have shown that the crack length a is an exponential function of the number of cycles N . Weakening of the grain boundaries leads to loss of intergranular adhesion. Increasing the yield stress by applying alloying materials. which will prohibit crack initiation and growth. 12. Besides the external surface.128 12. This phenomenon is referred to as creep fatigue. This process is called corrosion fatigue. 12. are also inﬂuencing fatigue life. At the external surface of the specimen.
typically < 0. which relate da the crack grow rate dN to load amplitude or maximum load.1 Crack growth models To predict the fatigue life of structures. such as the spacing between the striations. crack growth models have been proposed.5. crack growth is provoked by shear stresses and involves slip on a single crystallographic slip plane. The initial fatigue crack length ai seems to be a very important parameter for the fatigue life Nf . In this stage I. where a relative short number of cycles will result in fast crack growth leading to failure. where stresses are low. which can be detected by nondestructive techniques. because we assume to be in the high cycle fatigue regime.15 : Crack length increase with number of cycles. The crack growth is now much faster and after the fatigue life Nf its length is af and after a few cycles ac the critical crack length is reached and failure occurs. Nr N =1− Nf Nf 12.129 fatigue life. At this fatigue crack initiation life the initial crack has been formed. After Ni cycles. resulting in striations. in stage II of crack growth. the crack growth will be faster. For an initially undamaged material. Microstructural models relate the crack grow rate to microstructural parameters. the cycles to go until failure at Nf .25 nm/cycle. which can be expressed in the stress intensity factor K. All these relations describe dN to be a function of the stress σ and the . but in most cases it is so small that it cannot be detected. The crack propagation rate is very low. crack propagation is faster. typically µm’s per cycle. 12. The crack growth is provoked by tensile stresses and involves plastic slip on multiple slip planes at the crack tip. The restlife is the ratio of Nr and Nf . Phenomenological models are formulated and ﬁtted by close inspection of experimental da crack growth data. is indicated as Nr . Bates and Clark [3] found that for certain materials this spacing is about 6 times (∆K/E)2 . For higher loading amplitudes. After a large number of cycles the crack reaches a length a1 . a I II ac σ III ac af a1 ai Ni Nf N Fig. it takes Ni cycles to initiate a crack by dislocation movement and void coalescence. After N cycles.
For still smaller values of ∆K. in LEFM.7 → C = 2 × 10−9 dN → log [mm] √ m [MPa m] (−2) − (−4) =4 (2) − (1. 12. a crack grows extremely slowly. load and loading frequency. where scaling da with E improves the generality of the ﬁtted results. This is generally written as a relation between dN and ∆K. hampered by the roughness of the crack faces. C and m. For high values of ∆K. the Paris law is represented by a straight line. On a doublelogarithmic axes. . the crack growth is extremely small but not completely zero. However. as is shown in the example below.7 Fig. the lower limit. da = C(∆K)m dN 12.5) Limits of Paris law For low and high values of ∆K.16 : Paris law parameter ﬁt. when two data points are known. crack growth is much faster than predicted by the Paris law. which has become known as the Paris law in 1963 and it is still widely used [51]. da log dN [mm/c] 1 2 3 4 5 0 6 7 8 9 1 2 3 4 log(∆K) √ [MPa m] log C = −8.2 Paris law Paris & Erdogan published the crack grow law. using two parameters : a coeﬃcient and an exponent. da = C(∆K)m dN log(∆K) = 0 m= → da = log(C) + m log(∆K) dN da log(C) = log = −8. can be ﬁtted easily. dN is mostly related to ∆K by the socalled Paris law.5. Their values depend on material. geometry. The two parameters. Paris law will not describe the crack growth rate accurately any more. For ∆K ≈ Kth .130 da crack length a.
19 0. the conversion of units must be done with care.86 [ MPa ] 0.0254 [ m ] }m → [mm] and [MPa] 103 [ mm ] [m] √ m = √ = √ [ MPa m ] {[ MPa ] 103 [ mm ]}m [m] and [MPa] 1 103 √ { 103 }m [ mm ] √ [ MPa mm ]m .11 1.6 5.09)m [m] √ [ MPa m ]m da dN √ (∆σ πa)m → C= 0. Paris law parameters For some materials values of Paris law parameters are listed.24 0.6 4.2 3. The example shows the conversion from BS to SI units and vice versa.4.8 Conversion Because the dimension of C depends on the value of m.0.3 2.34 68.3. √ da = C (∆σ πa)m dN [in] and [ksi] 1 → [m] and [MPa] 0.3 3.0 1. 12.56 3 .85 . material mild steel structural steel idem in sea water aluminium aluminium alloy copper titanium ∆Kth [MNm−3/2 ] 3.0 2.0 1.0 1.0 2.0 m[] 3.8 2.07 .0 6.4 C×10−11 [!] 0.6 .131 da log( dN ) R rapid crack growth power law growth ∆Kth slow crack growth ∆Kc log(∆K) Fig.17 : Limits of Paris law.9 3.9 2.8 3.0254 [ m ] [ in ] √ = = [ ksi in ]m {6.0 1.5 2.0 1.9 4.2 2.0254 (1.
here. ∆N. (∆σ)−m Nf − Ni = m √ m β C( π) (1 − (1− m ) 2 m 2) (1− m ) af 2 1− ai af Paris law can also be integrated numerically.3 Fatigue life Paris law can be integrated analytically. m = 2. .56e−11 . m = 3. The initial crack length is indicated to be a0 and the ﬁnal crack length is the critical crack length ac .132 12.1 [mm] 80 a [mm] 60 40 20 0 0 1 2 N [c] 3 4 x 10 5 6 2 4 6 N [c] 8 10 12 14 x 10 6 aluminum .5. a = a0 while a < ac √ ∆K = β ∆σ π ∗ a da = C ∗ (∆K)m → dN a = a + ∆a N = N + ∆N end ∆a = da ∗ ∆N dN Initial crack length It is now very easy to show that the initial crack length a0 is of utmost importance for the fatigue life. ∆σ = 50 [MPa] Fig. C = 4. 12. set ∆σ.25e−11 .9 . ∆σ = 50 [MPa] mild steel . ac initialize N = 0. DN = 1000 120 a0 = 1 [mm] 100 80 a [mm] 60 40 20 0 0 a0 = 1. Results are shown for aluminum and for mild steel for a stress amplitude ∆σ = 50 MPa. where a increases from ai to af while N goes from Ni to Nf .1 [mm] C = 0.18 : Fatigue crack length in aluminum and mild steel for two values of the initial crack length. DN = 100 120 100 a0 = 1 [mm] a0 = 0. using an explicit integration scheme. The result Nf − Ni can be represented as a function of af with ai as parameter or vice versa.3 .
5. Erdogan (1963) da C(1 + β)m (∆K − ∆Kth )n = dN KIc − (1 + β)∆K Broek & Schijve (1963) da 2 = CKmax ∆K dN Forman (1967) da C(∆K)n = dN (1 − R)Kc − ∆K Donahue (1972) da = C(∆K − ∆Kth )m dN with ∆Kth = (1 − R)γ ∆Kth (R = 0) with R= Kmin Kmax with β= Kmax + Kmin Kmax − Kmin .1 7 572000 1 2 N [c] 3 x 10 4 7 Fig.5 1366000 100 0. 12.4 Other crack grow laws Numerous special and extended versions of the basic Paris law have been published. the number of cycles to reach this length has been calculated for various stress amplitudes. ∆σ [MPa] a0 [mm] ac [mm] Nf [c] 60 50 40 a [mm] 30 20 10 0 0 25 0. 12. when the crack length becomes critical (a = ac = 2γ ∆σ2 ). For π aluminum.19 : Fatigue crack length in aluminum for various load amplitudes.1 56 35070000 50 0.133 Fatigue load E The fatigue life is reached.1 12.1 28 4610000 75 0.
134 Walker (1970) da =C dN Priddle (1976) da =C dN ∆K − ∆Kth m KIc − Kmax with ∆Kth = A(1 − R)γ ∆K (1 − R)n
m
with
m = 0.4
;
n = 0.5
and
1 2
≤γ≤1
[Schijve (1979)]
McEvily & Gr¨ger (1977) o da A ∆K = (∆K − ∆Kth )2 1 + dN Eσv KIc − Kmax with ∆Kth = ∼ A, ∆K0 NASA / FLAGRO program (1989) C(1 − R)m ∆K n (∆K − ∆Kth )p da = dN [(1 − R)KIc − ∆K]q m=p=q=0 → Paris → Forman → Walker
1−R ∆K0 1+R inﬂuence environment
m = p = 0, q = 1
p = q = 0, m = (mw − 1)n
12.5.5
Crack growth at low cycle fatigue
Paris law and its extended version are valid for the high cycle fatigue regime, where stresses are so low that ∆K characterizes the stress amplitude. For low cycle fatigue this is not the case any more. Crack grow laws for this high stress regime with large plastic crack tip regions, are still under development.
σ λσ
slipline
θ θ
Fig. 12.20 : Crack under multimode fatigue loading.
135
1 3 − sin−2 (θ) cos−2 ( θ ) K da 2 = 1 − βγ − 2 dN 9 sin(θ) Eσv θ = cos−1 1 3 β → √ = 0.5 + 0.1R + 0.4R2 γ
2 Kmax {1 − (1 − λ) σmax } σv
2 K 7 Kmax da = √ (1 − 0.2R − 0.8R2 ) dN {1 − (1 − λ) σmax } 64 2 Eσv σv
For high values of crack tip stress and consequently a large plastic zone, the Paris law may be used with ∆J instead of ∆K. Note that ∆J = Jmax − Jmin over a cycle.
∗ da = C ∗ (∆ J)m dN
with
∆J =
Γ
∂∆ui W ∗ n1 − ∆ti ∂x1
εpqmax
dΓ
;
W∗ =
εpqmin
∆σij dεij
12.6
Load spectrum
Until now, we only looked at harmonic loads with constant frequency and amplitude. In practical situations, such loads will almost never be seen. Both loading frequency and amplitude will generally vary over time. Cyclic loading with such a time dependent character is referred to as a load spectrum. It is assumed here that the direction of the load remains the same. A load spectrum is shown in the ﬁgure below. It consists of four (i = 1..4) harmonic loads with diﬀerent frequencies and amplitudes. In the subsequent loads the number of cycles is ni . For each individual harmonic load the life time Nif is known. Miner has found in 1945 that the fatigue life for the whole spectrum is independent on the actual sequence of the individual loadings [47]. This means that the individual harmonic loadings do not inﬂuence the damage growth in the following loading stage with a diﬀerent ni character. The ﬁnal life time is reached, when the summation of Nif becomes one. This socalled Miner’s rule is approximately valid for a wide range of materials. In reality the ultimate sum of all the ratios is between 0.6 and 2.0, which indicates that there is actual some interaction between the diﬀerent load cycles.
136 σ
0 N
n1
n2
n3
n4
Fig. 12.21 : Sequence of various harmonic loadings.
L i=1
ni =1 Nif
Miner’s rule is clearly illustrated by regarding the rest life after each load cycle. When it ni reaches zero, the summation of the ratio Nif must reach the value one. n1 N1f n2 n1 − 1− N1f N2f n2 n1 n3 − 1− − N1f N2f N3f n1 n2 n3 n4 1− − − =0 − N1f N2f N3f N4f
1 2 3 4
→ → → →
1−
12.6.1
Random load
Although the load spectrum is already much more realistic than the harmonic loading with constant frequency and amplitude, practical loading is mostly rather random, as is illustrated in the ﬁgure below. Prediction of fatigue life is only possible after this random load is transferred into a load spectrum, a sequence of harmonic loads, with known frequencies and amplitudes. Various counting procedures can be used for this transformation. Literature gives detailed information about (mean crossing) peak count, range pair (mean) count and rain ﬂow count techniques.
22 : Random load. random load sequences are measured and used in design. 12. For some applications. During service life the load can be monitored and the expected fatigue life time can be updated. a sea vessel (b) and an airplane (c) are shown in the ﬁgure.137 σ 0 t Fig. Typical loadings for a manufacturing machine (a). .
The question is how they aﬀect the fatigue life. 12. Although the intuitive feeling is that overload must be bad. where (Kmax )OL is the maximum SIF during the overload. at least during a limited number of cycles. which must of course be low enough to cause no direct damage. During service life extreme overloads can occur. .138 Fig.2 Tensile overload Known random loadings are used to calculate the fatigue life during the design of structures. sea vessel (b) and airplane (c). practical experience shows that the crack propagation is slowed down by such an overload.23 : Random loads for machine (a). 12.6. This is illustrated in the next ﬁgure.
12.5 16. and . In a certain time period. the loading stress is increased to a maximum value σ1 .5 16. The ﬁgure shows the stress σyy and the plastic limit stress.5 22 44 9 55 245 The crack retardation can be explained by regarding the stress state in a region in front of the crack tip. The crack retardation is also illustrated quantitatively by the values in the next table (from [29]).1 23.139 (Kmax )OL Kmax a a b2 b1 N Fig.5 16.9 10.5 23.5 16. representing the overload. ∆K √ [MPa m] 15 15 15 16.1 % Pmax [] 53 82 109 50 50 50 50 50 50 50 75 100 nr.24 : Overload and crack retardation.5 16.1 23. which apply to the Mode I loading of specimen made of Al 2024T3 alloy. Pmax [] 1 1 1 1 10 100 450 2000 9000 1 1 1 delay [103 cycles] 6 16 59 4 5 9.
When the crack tip has moved through the compressive zone. Both relations are taking into account the inﬂuence of the compressive stress zone at the front of the crack tip.140 decreased again. σ σ σyy σv σ1 0 t σ A B σyy A1 B1 σyy σ=0 B2 εyy A2 Fig. the normal crack grow rate is again established. Crack retardation models Willenborg (1971) has published a relation for the SIF. KR . .25 : Residual stress at the crack tip resulting from overload. in the crack retardation stage. the strain in A and B is more or less the same. In two points A and B the stress σyy is plotted in the ﬁgure as a function of the strain εyy . In point A. After unloading. 12. so that it becomes clear from the stressstrain curves that a compressive residual stress remains in point A. It has been enhanced by Johnsson [37]. while the deformation in point B remains elastic. This compressive stress inhibits the opening of the crack and thus the crack elongation. the material is plastically deformed.
(Kmax )OL Kmax a ∆a ry Johnson (1981) R ef f Kmin − KR 1 (Kmax )OL = . Elber (1971) [17] Schijve (1981) [58] ∆Kef f = U ∆K . socalled design against fatigue. the stress amplitude in the most severely loaded parts of the structure are kept below the fatigue limit.55 + 0. The result is that fatigue crack growth will be practically zero.141 Willenborg (1971) KR = φ (Kmax )OL 1− ∆a ry . ry = Kmax − KR βπ σv β = plastic constraint factor 2 Crack retardation due to crack closure.7 Design against fatigue There are three diﬀerent strategies which can be followed to prevent that fatigue damage leads to failure.7 U = 0.1 ≤ R ≤ 0. U = 0. caused by the compressive stress zone behind the crack tip.4R with with − 0.33R + 0. This strategy will involve the use of much material and thus lead to bulky and heavy structures. − Kmax KR = 0 −1 . Inﬁnite life design For inﬁnite life design.12R2 − 1.54 12. have also been devised and published.5 + 0.0 < R < 0. → ∆a < ry delay distance S = shutoﬀ ratio KR = residual SIF φ = [1 − (Kth /Kmax )](S − 1) . .
the structur(e)(al) (part) is replaced. When the load is random. They can be combined in a stress/strain life curve. Instead of the (SN)curve. When the number of cycles to failure – probable a much lower safe number – is reached. including eﬀects like overloads. The rest life can be calculated. using a crack growth relation like Paris law. satistical counting procedures must be employed to translate the random load spectrum to a sequence of harmonic loads. The design must be such that (undetected) fatigue cracks will not lead to global failure. the structur(e)(al) (part) is replaced. When the critical crack length – probable a much smaller value – is reached. Periodic inspection is of course needed to check whether predicted crack lengths are correct. Damage tolerant design Crack growth can be predicted. the (SN)curve is used to predict when the fatigue life is reached. the loading must be known and the number of cycles must be logged. For this purpose. . Fail safe design Although safety factors will certainly be used to be within safe margins of the fatigue life. the ﬁtted Basquin or MansonCoﬃn relations can be used.142 Safe life design Using the concept of safe life.
Chapter 13 Engineering plastics (polymers) Polymer materials are widely used in functional and structural applications. External loads and other inﬂuences lead of course to damage. cracking and failure. There are a huge number of diﬀerent polymer materials. has given rise to the development of new materials with better damage resistance. even in the linear elastic case. Temperature eﬀects cannot be neglected in many cases. Composite materials are anisotropic. in which case Linear Elastic Fracture Mechanics can be used to study cracks. Polymers mostly have a lower stiﬀness and the linear elasticity limit is reached sooner. complicates the mechanical behavior. even more than in the case of metal alloys. which.1 Mechanical properties For many applications. Polymer materials generally show strainrate and timedependent behavior. metal alloys show linear elastic behavior. 143 . some of which are listed in the table. In fracture analysis of polymers it is therefore often needed to use concepts from NonLinear Fracture Mechanics. ABS HIPS LDPE Nylon PC PMMA PP PPO PS PSF PTFE PVC PVF PVF2 acrylonitrilbutadieenstyreen highimpact polystyreen lowdensity polyetheen polycarbonaat polymethylmethacrylaat (plexiglas) polypropeen polyfenyleenoxyde polystyreen PolySulFon polytetraﬂuoretheen (teﬂon) polyvinylchloride polyvinylﬂuoride polyvinylidieenﬂuoride EM TP TP TP TP TP TP TP TP TP TP TP TP 13. Systematic study of these phenomena has brought insight in their behavior and.
13.1. In a craze. ﬁbrils are connecting the two craze surfaces. In this case no change in density is observed. . These can occur both in the same material in regions with diﬀerent stress states. the ﬁbril snaps. because crazing is generally associated to brittle behavior. because crystals can stop cracks in an early stage. When the maximum elongation is reached. The local density may decrease with 40 . 13. When the craze opens. Crazes are connected to normal yielding. shear bands crazes Fig.2 Properties of engineering plastics The next table shows some properties of engeneering plastics. which are related to their damage behavior. The crystallinity is a very important item.1 : Shear bands and crazes in a polymer.60 %.144 13. Shear bands are associated to shear yielding.1.1 Damage Damage in polymers may occur as shearbands or crazes. The occurence of crazes is also relevant. the ﬁbrils are elongated due to extrusion of macromolecules from the bulk polymer.
.145 PMMA PS PSF PC Nylon 66 AC a a a a cr CZ + +  KIc 13. main chain mobility and crystallinity.2 17. Blending one polymer with another polymer or particles to improve ductility and fatigue life is referred to as toughening.6 low high PVF2 PET CPLS HIPS ABS AC : AC : AC : AC : CZ KIc a c sc cr sc sc + main chain segmental motions → energy dissipation main chain segmental motions → energy dissipation crystalline regions → crack retardation amorphous → strain induced crystallization at crack tip crosslinked → suppressed crazing µsized rubber spheres → enhanced crazing blending = = = = = = amorphous crystalline semicrystalline crystalline regions crazing √ fracture toughness in MPa m 13. cross linking. Data for the next ﬁgures are taken from [29].2 Fatigue Fatigue crack propagation in polymers is strongly aﬀected by properties like molecular weight.
6 PVF2 PET .0 10.0 √ ∆K [MPa m] 1 2 3 4 9 : : : : : PS PMMA PSF PPO PVC 5 6 7 8 : : : : PC Nylon 6.146 10 da dN [mm/cycle] 3 ∗ 10−3 10−3 −2 4 2 3 5 1 3 ∗ 10−4 10−4 0.5 1.0 2.0 5.5 1 2 5 √ ∆K [MPa m] crazing 1 2 3 4 5 : : : : : PMMA LDPE ABS PC Nylon 5 Hz 1 Hz 10 Hz 10 Hz 10 Hz 10−1 da dN [mm/cycle] 10−2 no crazing cristalline regions 2 3 4 5 6 7 10−3 8 10−4 1 9 10−5 0.
147 da dN [mm/cycle] 10−2 2 1 3 4 10−3 10−4 0.0 2.0 5.0 √ ∆K [MPa m] .5 1 2 3 4 : : : : CLPS PS HIPS ABS 1.
148 .
with special attention for their fatigue life. so the cohesive zone law comprises the two relations Tn (∆n ) and Tt (∆t ). Uncoupled laws are intended to be used when the debonding process occurs under one mode – normal (modeI. leading to a constitutive relation. Also the peeloﬀ lid of food containers is relying on the proper delamination of a polymer interface layer. These models are used to investigate the behavior of interfaces between polymer coating and steel sheet substrate. Interaction between crack faces is automatically incorporated and can be ﬁtted on experimental data.1 Cohesive zone models Cohesive zone (cz) models have been introduced by Dugdale and Barenblatt [15] and have attracted a growing interest in the scientiﬁc community to describe failure processes and delamination in particular. which is achieved by introducing coupling parameters in the model. who’s papers and reports are listed in the reference u c section []. As a ﬁnal example. both normal and tangential tractions depend on both the normal and tangential opening displacement. Frequently – but not necessarily – a diﬀerence is made between normal (n) and tangential (t) direction. Cohesive zone models relate the relative displacement (”opening” ∆) of two associated points of the interface to the force per unit of area (”traction” T ) needed for separation. opening) or tangential (modeII. interfaces get more important concerning the (mechanical) behavior of material systems. Cohesive zones project all damage mechanisms in and around a crack tip on the interface. The next sections are based on the research work of Marco van den Bosch (PhD). 149 . Sil Bijker (MSc) and M¨ge Erin¸ (PhD). The majority of cohesive zone laws have a (partial) coupling between normal and tangential direction. the interfaces in solder joints for microelectronic components are studied. sliding) loading – or is largely dominated by one mode. In a coupled cohesive zone law.Chapter 14 Cohesive zones With ongoing miniaturization. The degradation of interfaces – e. a material system used more and more for all kind of applications. or cohesive zone law.g. delamination – can be studied and analyzed with cohesive zone models. Cohesive zone laws can be uncoupled or coupled. 14. In an uncoupled cohesive zone law the normal/tangential traction is independent of the tangential/normal opening.
The ﬁgure below illustrates this cohesive zone behavior with the function f (λ) taken from [64]. These cohesive zone laws are described in the next sections.max Tt = Tt. 0 < λ ≤ 1) δt δt ∆n 2 ∆t 2 + δn δt 27 1 − 2λ + λ2 f (λ) = 4 λ= eﬀective displacement function from [64] .150 n t Fig. which is the same. 14.max f (λ. rigidlinear and exponential.1. traction = f(opening) normal / tangential coupling T = f (∆) Tn = fn (∆n ) . The largest tractions are associated with the opening in one direction or.1 : An undeformed and deformed interface with a schematic illustration of a cohesive zone tractionopening interaction in normal and tangential direction. with an uncoupled formulation. Tt = ft (∆t ) Tt = ft (∆t . Note that in this cohesive zone δn and δt indicate the opening. The maximum normal traction and the maximum tangential traction are indicated by Tn. The areas below the tractionopening curves represent the normal and tangential workofseparation Wn and Wt . The coupling is incorporated well. respectively.max ∆n f (λ. The plots show that the tractions in one direction deminish. The function f (λ) is such that the tractions are nonzero for 0 < λ ≤ 1 and remain zero for λ > 1. Tn = fn (∆n .1 Polynomial In the polynomial cohesive zone law. 14. coupling is achieved with an eﬀective opening displacement λ and a coupling parameter α. piecewise linear. which relates the maximum tractions.max . Most of them can be categorized into the following groups: polynomial. when the traction is zero.max and Tt. 0 < λ ≤ 1) = αTn. ∆n ) A large variety of cohesive zone laws has been described in literature. Tn = Tn. 0 < λ ≤ 1) δn ∆t ∆t f (λ. Characteristic openings are δn and δt . when the cohesive zone is opened in the other direction. ∆t ) .
5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig.1. 14. .3 : Piecewise linear cohesive zone law without coupling. shown below. 14. the normal traction is not a function of the tangential opening ∆t . Due to the assumed relation between the maximum tractions.1.5 −0.max = (δn /δt )Tn.5 Tn /Tn. Tt. the normal and tangential workofseparation have the same value.5 Tn /Tn. 14.5 0 0 −0.max 1 1 0.2 Piecewise linear In the piecewise linear cohesive zone law. This implies that the interface can only fail due to decohesion in normal direction.5 0 0 −0. The plot shows a rigidlinear law without this coupling (α = 1).5 −0.5 −1 −1 0 1 ∆n /δn −1 2 3 4 −2 −1 0 ∆t /δt 1 2 Fig.max 0. No coupling is taken into account here. until the maximum traction is reached. In such a rigidlinear cohesive zone law.max 0. the tractionopening function is trapezoidal.2 : Polynomial cohesive zone law with coupling.max Tt /Tt.max Tt /Tt.3 Rigidlinear Cohesive zones may be modelled to have a very high stiﬀness. The characteristic openings δn and δt are the openings at zero traction. while the tangential traction is controlled completely by the normal opening ∆n . 14. as proposed in [12].151 1 1 0.
r.5 −1 −1 0 1 ∆n /δn −1 2 3 4 −2 −1 0 ∆t /δt 1 2 Fig. but appears to have some deﬁciancies in describing coupling. which will be discussed below.5 −0. It has some advantages compared to other laws.4 : Rigidlinear cohesive zone law without coupling. proposed by Xu & Needleman is much used. the tractions and their derivatives are continuous.5 Tn /Tn. which is attractive from a computational point of view. ∆n and ∆t respectively. 14.1.max ∆n δn ∆n > 0 1 0. The exponential cohesive zone law originates from the universal relationship between binding energies and atomic separation of interfaces. both based on a potential φ. 14. First of all. which is a function of both the normal and the tangential opening. Secondly.152 Tn = Tn.max ∆t 1−α δt α= ∆t > 0 δt ∆n ∆t δn coupling no coupling α=1 1 0. a phenomenological description of contact is automatically achieved in normal compression.t.max 1 − Tt = Tt. q and r. Two exponential czlaws are described in the following.max 0 0 −0. There are two coupling parameters. ∆n δn 1−q r−1 φ(∆n .4 Exponential The exponential cohesive zone law is used very much. Xu & Needlemaan exponential czlaw The Xu&Needleman cohesive zone law [69] is deﬁned by one potential φ which depends on the normal and tangential openings ∆n and ∆t . The ﬁrst law. ∆t ) = φn + φn exp − 1−r+ ∆n δn − q+ r−q r−1 ∆n δn exp − ∆2 t δt 2 Tractions in normal (Tn ) and tangential (Tt ) direction are derived from the potential φ by diﬀerentiation w. .5 Tt /Tt.
the tractionopening relations are simpliﬁed. ∆n φn ∆n exp − δn δn δn φn ∆t ∆2 t Tt = Tt (∆n = 0) = 2 q exp − 2 δt δt δt Tn = Tn (∆t = 0) = =2 ∆2 φt ∆t t exp − 2 δt δt δt Tn. For r = q and other values of r.5 Tn /Tn. The normal and tangential tractionopening relations are plotted for the parameter values φn = 100 Jm−2 . Introducing the parameter φt .max = 10 N . the results are even more unrealistic.max exp(1) φt Tt. When the cohesive zone is loaded simultaneously (or subsequently) in normal and tangential direction. as the normal traction is not zero for zero normal opening. Tn.max 0. 14. For r = 0 the normal traction does not become zero. The results are dependent of the value of r. φt = 80 Jm−2 .5 0 0 −0.max = Tn (∆n = δn ) = Tt. Tt.5 : Uncoupled normal and tangential tractionopening for the Xu&Needleman cohesive zone law. the coupling parameter q is deﬁned as q = φt /φn .max = 10 N .max 1 2 exp(1) 1 0. the coupling is taken into account.max = Tt (∆t = φn δn exp(1) φt 1 2 δt ) = 1 δt 2 exp(1) 1 → → δn = δt = φn Tn. when the interface is failed in tangential direction. .153 Tn = ∆n φn ∆n ∆2 1−q ∂φ ∆2 t t = exp − exp − 2 + 1 − exp − 2 ∂∆n δn δn δn δt r−1 δt ∂φ r − q ∆n φn ∆t ∆n ∆2 t Tt = =2 q+ exp − exp − 2 ∂∆t δt δt r − 1 δn δn δt r− ∆n δn For pure normal and tangential loading.max Tt /Tt.5 −0.5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig.
max 0.t. For pure normal and tangential opening. is again based on a potential [5]. The potential incorporates four independent parameters: the workofseparation for pure normal opening. 14.5 Tn /Tn. 1 1 0.max Tt /Tt.5 −0. respectively.max Tt /Tt. Adapted exponential cohesive zone law An exponential cohesive zone law. 14.5 0 0 −0.154 1 1 0. the characteristic opening in normal direction. φ(∆n . φn .5 −0. the normal and tangential opening. δt . which does not have the anomalities of the Xu&Needleman law.5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig. The normal and tangential tractions are calculated by diﬀerentiating the potential w.5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig.7 : Coupled normal and tangential tractionopening for the Xu&Needleman cohesive zone law for r = q.5 Tn /Tn. φt . The characteristic lengths δn and δt can be expressed in the maximum tractions.5 0 0 −0.6 : Coupled normal and tangential tractionopening for the Xu&Needleman cohesive zone law for r = 0. δn . and the characteristic opening in tangential direction.max 0.r. the tractions are plotted in the ﬁgure below and they are identical to those of the Xu&Needleman law. the workofseparation for pure tangential opening. ∆t ) = φnt 1 − 1 + Tn = ∂φ ∂∆n = φn ∆n δn exp − ∆n δn ∆n δn exp − ∆2 t δt 2 ∆2 t δt 2 φn δn φt δt ∆n δn ∆t δt exp − 1+ exp − ∂φ Tt = ∂∆t =2 φt ∆n δn exp − ∆2 t δt 2 exp − ∆n δn .
5 Tn /Tn.g.5 −0.max at ∆t /δt = 1/ 2.5 Tn /Tn.max Tt /Tt.max 1 φt 1 2 exp(1) 1 0. right: The uncoupled tangential response of √ the exponential cohesive zone law with the maximum traction Tt.155 δn = φn Tn. is illustrated by the tractionopening plots below. in the case of a moving delamination front) and a cohesive zone completely fails in shear.5 0 0 −0.5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig. 14.8 : left: The uncoupled normal response of the exponential cohesive zone law with the maximum traction Tn. .5 −0. 1 1 0.max 0.max at ∆n /δn = 1.5 0 0 −0. If complete loss of interfacial integrity is important (e. its loadcarrying capacity in normal traction should completely vanish as well and vise versa.max Tt /Tt.9 : The coupled normal and tangential response of the exponential cohesive zone law.max exp(1) .max 0. That this behavior is described adequately. Coupling Adequate coupling between the normal and tangential directions is required in a cohesive zone law to describe the physically occurring interface behavior realistically. 14.5 −1 −1 0 1 ∆n /δn 2 3 4 −1 −2 −1 0 ∆t /δt 1 2 Fig. δt = Tt.
.max )} 1 ∗ Tn = max{Tn (∆n . ∆n.max . ∆t.6 0.max /δn Fig.8 0.max 0. The shaded areas under the curves are the work of separation for normal and tangential opening.8 ∗ Tn /Tn. the maximum shear traction decreases to zero for increasing normal separations. as is illustrated in the ﬁgure below.2 0. However.2 0 −1 0 1 2 3 4 0 −2 −1 0 ∆t. Theoretically.4 0. As can be seen in this ﬁgure.max d∆t Fig.10 : Loading sequence to study the inﬂuence of coupling parameters on workofseparation: (left) interface loaded to ∆n. this can also happen in compression. In the left ﬁgure below the evolution of the maximum shear traction is shown as a function of the normal separation. these negative values of ∆n cannot be reached in practice since the normal compressive traction increases very fast in this regime.max ∞ Wn = 0 Tn (∆n )∆t =0 d∆n . 14. right: The normalized maximum normal traction as a function of the normalized tangential separation of the alternative exponential law.4 0. 14.max .11 : left: The normalized maximum shear traction as a function of the normalized normal separation of the alternative exponential law.156 To investigate this behavior in more detail. (right) interface broken in shear ∆t → ∞. ∆n. Tt∗ = max{Tt (∆t . After that it is broken in shear : ∆t → ∞. The maximum normal traction decreases to zero for increasing tangential separation. Wt = 0 Tt (∆t )∆n =∆n.max )} 1 0. the interface is ﬁrst loaded in normal direction until a maximum opening ∆n.max /δt 1 2 ∆n.max ∗ Tt /Tt. Several loading sequences are evaluated. In the right ﬁgure below the maximum normal traction as a function of the tangential separation is shown.6 0.
In the ﬁgure below it can be seen that Wtot = φn when the cohesive zone is only loaded in normal direction (modeI: α = 0o ) and Wtot = φt when it is only loaded in tangential direction (modeII: α = 90o ). values φn = 100 Jm−2 and φt = 80 Jm−2 are used in these calculations.r.12 : left: Workofseparation when the cohesive zone is ﬁrst loaded in normal direction until ∆n. When material acceleration is suﬃciently small. Wtot increases monotonically from the value φt to the value of φn .max and then broken in tension. the normal direction until complete separation occurs. Wtot smoothly decreases from the value of φn to φt . Modemixity The inﬂuence of mixed mode loading on the total workofseparation is studied next. 14. a cohesive zone is loaded under an angle α w. and the corresponding value of Wtot is calculated. Again. when considering the total workofseparation Wtot = Wn + Wt in case of a sequential loading.2 Weighted residual formulation with cohesive zones In the deformed state of a material body. of equilibrium ∇ · σ + ρq = 0 ∀ x∈V . a partial diﬀerential equation.max and then complete normal separation.max and then broken in shear. the righthand side is zero and the equilibrium equation results. ˙ ∇ · σ + ρq = ρv ∀ x∈V eq. 14. right: Workofseparation when the cohesive zone is ﬁrst loaded in tangential direction until ∆t.157 Coupling should also be realistic. 100 100 80 80 W[Jm ] 40 W 20 W W 0 0 2 4 ∆n. the equations of motion must be satisﬁed in each material point with position x in the current material volume V . which can be solved with proper boundary conditions.max and then complete tangential separation. To this purpose.t.. in normal and tangential direction or viceversa.max /δt 6 8 Fig.max /δn n t tot W[Jm ] −2 −2 60 60 40 W 20 W W n t tot 6 8 0 0 2 4 ∆t. In the other case where there is ﬁrst tangential separation until ∆t. The behavior is as expected and shown in the ﬁgures below : in the case where there is ﬁrst normal separation until ∆n. Wtot is shown as a function of the loading angle α.
q) The internal load integral can be written as the sum of an integral over bulk material (volume Vb ) and an integral over cohesive zones (volume Vcz ). the weighted residual integral becomes the virtual work equation in which the logarithmic strain εln is introduced. We only consider the integral over the cohesive zones. Furthermore. fi = V (∇w)c : σ dV = Vb (∇w)c : σ dV + Vcz (∇w)c : σ dV = fib + ficz When the weighting function w is chosen to be a virtual displacement δu. The deformed opening of the cohesive zone is h. The unity vector e is deﬁned along the line between two associated points of the interface. t. where the cz has uniform depth d. we consider a twodimensional case. σ) = fe (w. ∇ · σ + ρq = 0 V ∀ x∈V V ⇐⇒ A V w · ∇ · σ + ρq ∀ w(x) dV = 0 ∀ w(x) (∇w)c : σ dV = w · ρq dV + ∀ w(x) w · t dA fi (w. The remaining integral is the internal load integral fi . (∇w)c : σ dV w = δu ficz = Vcz ficz = Vcz → (∇w)c : σ = (∇δu)c : σ = δεln : σ δεln : σ dA Acz δεln : σ dV = d ficz δεln = δεln ee = δεln e → δεln : σ = δεln · σ σ = σee = σe ∆(δu) δ∆u · σ dA = d · σ dA δεln · σ dA = d =d h Acz Acz h Acz d ∆w · σ dA = h Acz .158 with ∇ σ ρ q v : : : : : gradient operator Cauchy stress tensor density volume load material velocity Analytical solutions are seldomly reached for practical problems and we have to be satisﬁed with approximate solutions. The two integrals with the external loads are abbreviated as the external load integral fe . which can be determined with the ﬁnite element method. Partial integration of the ﬁrst term leads to a weak version. To this purpose the diﬀerential equation is transformed into a weighted residual integral by use of a vectorial weighting function w(x).
the weighted residual integral becomes the virtual work equation in which the linear strain is introduced. T ) = fe0 (w. q) ∇ = F −c · ∇0 → ∀ w(x) V dV = det(F ) dV0 = J dV0 (∇w)c = (∇0 w)c · F −1 ∀ ∀ w(x) w(x) V0 (∇0 w)c · F −1 : σJ dV0 = fe0 (∇0 w)c : (F −1 · σJ) dV0 = fe0 (∇0 w)c : T dV0 = fe0 ∀ ∀ V0 w(x) w(x) V0 fi0 (w. The unity vector e0 is deﬁned along the line between two associated points of the interface. t. In this transformation the deformation tensor F is used. In this case we choose the initial undeformed state. q0 ) Again the internal load integral is written as sum of an integral over the bulk volume and an integral over the cohesive zones.159 When large deformations occur. the integral has to be transformed to a known state. The transformed integral contains the ﬁrst PiolaKirchhoﬀ stress tensor T . fi0 = V0 (∇0 w)c : T dV0 = V0b (∇0 w)c : T dV0 + V0cz (∇0 w)c : T dV0 = fi0b + fi0cz When the weighting function w is chosen to be a virtual displacement δu. The undeformed opening of the cohesive zone is h0 . t0 . fi0cz = V0cz (∇0 w)c : T dV0 w = δu → (∇0 w)c : T = (∇0 δu)c : T = δεl : T δεl : T dA0 A0cz fi0cz = δεl : T dV0 = d0 V0cz δεl = δεl e0 e0 = δεl e0 T = T e0 e0 = T e0 fi0cz = d0 d0 = h0 δεl · T dA0 = d0 A0cz A0cz → δεl : T = δεl · T ∆(δu) · T dA0 h0 δ∆u · T dA0 = d0 h0 A0cz ∆w · T dA0 A0cz . We only consider the integral over the cohesive zones and take the cz to have a uniform initial depth d0 . (∇w)c : σ dV = fe (w.
Along P Q the local coordinate ξ is deﬁned in a range [−1.3 Twodimensional CZ element A cohesive zone element has uniform initial thickness d0 . 3 4 A 1 et en = ∆wT e ˜ ˜ T T en et 2 B ∆w = ∆wt et + ∆wn en = [∆wt ∆wn ] T = Tt et + Tn en = fi0cz = d0 l0 2 1 Tt Tn et en = TTe ˜ ˜ ∆wT (η)T (η) dη ˜ ˜ η=−1 . a local orthonormal basis {et .160 14. This traction is a function of the elongation of the line P Q w. Along this line a local coordinate η is introduced. which spans the range [−1.1 Local vector base The weighting function and the traction are written in components w. 3 Q 4 A 1 1 P d0 h0 1 ξ=−1 1 T ξ 0 η 1 B 2 fi0cz = = d0 h0 d0 2 ∆w · T dA0 = A0cz 1 ξ=−1 1 ∆w · T η=−1 1 h0 l0 dηdξ 2 2 ∆w · T η=−1 d0 l0 l0 dηdξ = 2 2 ∆w · T dη η=−1 14. 1].3. The length l0 of the element is deﬁned to be the undeformed length of the line AB between the midpoints of the element edges 14 and 23.r.t. which is deﬁned in the midpoint of AB. the undeformed state.r. In the deformed state a traction T works between associated (= with the same ηcoordinate) points P and Q on edges 12 and 43 respectively. en }.t. given by the cohesive zone law. 1].
which is done with a rotation matrix R. en } to the global base vectors {ex . . The interpolation functions are linear in the local coordinate.t. The only nonzero components in the rotation matrix are cosine (c) and sine (s) functions of the angle between the line AB and the global xaxis.r. It transforms the local base vectors {et . 1 0 2 (1 − η) 1 0 A A B B 2 (1 − η) ∆wT (η) = [∆wt ∆wn ] == ∆wt ∆wn ∆wt ∆wn 1 (1 + η) 0 ˜ 2 1 0 2 (1 + η) T = ∆w AB N T (η) ˜ The values ∆w in points A and B are written as the diﬀerences of the weighting function in the nodal points 4 and 1 and 3 and 2 respectively. en }. The cosine and sine values can be easily calculated from the coordinates of points A and B : c = (xB − xA )/ℓ and s = (y B − y A )/ℓ. ˜ 1 wt w1 n 2 A −1 0 0 0 0 0 1 0 wt ∆wt ∆wA 0 −1 0 0 0 0 0 1 w2 n AB n = ∆w B = 0 0 −1 0 1 0 0 0 w3 = P w l t ∆wt ˜ ˜ 3 B ∆wn 0 0 0 −1 0 1 0 0 wn 4 wt 4 wn The column with weighting functions is transformed from the local to the global coordinate system. the local vector basis {et . 2 1 wt 1 wn 2 wt 2 wn 3 wt 3 wn 4 wt 4 wn wl = ˜ = c −s 0 0 0 0 0 0 s 0 c 0 0 c 0 −s 0 0 0 0 0 0 0 0 0 0 0 0 s 0 c 0 0 c 0 −s 0 0 0 0 0 0 0 0 0 0 0 0 s 0 c 0 0 c 0 −s 0 0 0 0 0 0 s c 1 w1 1 w2 2 w1 2 w2 3 w1 3 w2 4 w1 4 w2 = Rw ˜ . where ℓ is the deformed length of line AB. A 4 1 ∆wn = wn − wn B 3 2 ∆wn = wn − wn A matrix P is introduced to express ∆w AB in the local nodal displacements in the column ˜ wl . The index l indicates that components are taken w. ey }. Coordinates of A and B can be expressed in the nodal point coordinates : xy AB = 1 (xy 12 + xy 43 ).161 Interpolation The weighting function components are interpolated between their values in point A and point B. B 3 2 ∆wt = wt − wt A 4 1 ∆wt = wt − wt .
Also the rotation of the local to the global coordinate system is applied. which means that each unknown quantity is written as its approximated value ( )∗ and an iterative change δ( ). T ) = f e0 ˜ ˜ ˜ b ˜ cz w ˜ 14. This means that the nodal point internal forces have to be in equilibrium with the external nodal forces. T ) ˜ ˜ ˜ ˜ cz η=−1 1 After assembling all elements.162 So we can write : ∆wT = wT P T N T (η) = wT RT P T N T (η) l ˜ ˜ ˜ and fi0cz = w T ˜ d0 ℓ0 T T R P 2 N T (η)T (η) dη = w T f i0 (R. R = R∗ + δR ≈ R∗ d0 ℓ0 ∗T T R P 2 1 → f i = f ∗ + δf i i ˜ ˜ ˜ δf i0 = ˜ cz N T (η) δT (η) dη ˜ η=−1 The relation between the iterative traction and the iterative opening represents the stiﬀness of the cohesive zone. which is linearized subsequently. The iterative internal load column can now be expressed in the iterative opening. The iterative opening is expressed in the iterative nodal displacements in the global directions. The iterative tractions are then expressed in the iterative nodal displacements.3. T = T ∗ + δT ˜ ˜ ˜ . The unknown opening of the cohesive zone must be solved from nonlinear (global) equations in an iterative procedure. This is substituted in the internal load column. cz . The iterative changes of the traction can be expressed in the iterative opening. ∀ wT f i0 + w T f i0 = wT f e0 ˜ ˜ ˜ ˜ b ˜ ˜ cz f i0 (σ. It is assumed here that δR ≈ 0. the global weighted residual equation has to be satisﬁed for all nodal values of the weighting function. F ) + f i0 (R. using the stiﬀness matrix K ∗ of the cohesive zone element. ∂Tt ∂Tt δ∆t ∂T δT (η) = ˜ δ∆ = ∂∆t ∂∆n ∂T ∂T δ∆n = M (η) δ∆(η) ∂∆ ˜ ˜ ˜ n n ˜ ∂∆t ∂∆n The iterative opening is now interpolated analoguously to the interpolation of the weighting function.2 Iterative procedure The internal load column f i0 is a nonlinear function of the traction T and the element ˜ ˜ cz rotation R.
2 0. In the next example the cohesive zone is opened in normal and tangential direction simultaneously without rotation.5 Tt /Tt. Tangential tractions in the integration points of the czelement diﬀer also.1 0.5 1 2 −1 −0. 8 7 6 1.5 1 1. 14.8 ip1 ip2 0.max 0.5 −0.5 2 2.2 −0. The plots show the reacton forces in x− and y−direction in the upper nodes.1 3 4 1 2 1 2 3 4 5 0.05 0.max 1 2 3 4 5 5 6 fy /Tn.6 0.2 0 −0.5 1 1. 14.3 0.8 0.4 0.3 −0.max Tt /Tt. The ﬁgure below shows the opening and sliding mode loading.5 2 2.15 0. where the displacements of the upper nodes are prescribed.14 : Mixed mode opening of a cohesive zone element without rotation.5 1 0.05 0 0 fy /Tn.5 1 1.5 0 3 4 5 6 0 1 2 3 4 −0.max 5 6 0.2 0 0 0 3 4 −0. The cohesive zone is now opened in normal direction with a diﬀerent displacement in the upper nodes.25 0.4 0.2 0.max np7 np8 ip1 ip2 0.1 0 0. which have linear elastic material behavior.15 0. 14.6 0.5 7 8 fx /Tt.13 : Normal and tangential opening of a cohesive zone element. The forces are the same as the traction opening relaton of the cohesive zone.max 1.max fx /Tt.8 0.2 0 0.5 0 0.max np7 np8 ip1 ip2 0.4 0.5 0.4 1.7 0.max 0.5 0.4 0.2 0 0 0. the normal and tangential tractions do not reach their maximum values.5 2 2.2 0 0 0.1 0.4 −0.max 1 0. 7 8 1 np7 np8 1 np7 np8 fy /Tn. Reaction forces in xdirection are zero.15 : Opening of a cohesive zone element with rotation.6 0. Due to the coupling in the tractionopening law.5 1 1.5 −0. resulting in rotation.2 uy /δn ux /δt ∆n /δn ∆t /δt Fig.5 uy /δn ux /δt ∆n /δn ∆t /δt Fig.6 ip1 ip2 fx /Tt.8 0.5 5 0.5 0 0 0.163 Examples The cohesive zone element is placed between two bulk elements.5 2 2.6 0.5 0 −0.2 1 np7 np8 1 Tn /Tn.5 1 2 −1 −1 0 1 2 3 4 −1 −3 −2 −1 0 1 2 3 uy /δn ux /δt Fig. The plots show reaction forces in the upper nodes and the tractionopening behavior in the integration points of the cohesive zone element.4 0. . 7 8 2 1 np7 np8 1 Tn /Tn.
8 8 8 7 6 7 6 7 6 0.04 0.35 ip1 ip2 0.1 0.02 0. The characteristic opening length δ is the opening for which T reaches the maximum value Tmax .04 −0. Instead of deﬁning two separate constitutive relations for the normal and tangential direction.4 CZ for large deformation For large deformation it is no longer physical to discriminate between normal and tangential openings.15 0.01 −0.03 −0. .2 0.06 −0. T = f (∆) Tmax = .164 For large rotations it is obvious that the normal and tangential directions in the cohesive zone element are changing during the opening.2 0 0 ∆t /δt ∆t /δt ∆t /δt Fig. T = T (∆)e . whereby no distinction will be made between normal and tangential loadings. In the following three analyses the reference plane is chosen subsequently in the midplane (line AB). 40 30 T [MPa] 20 10 0 0 1 2 3 ∆ [µm] 4 5 6 Fig.1 0.03 ip1 ip2 Tt /Tt. The ﬁgure shows the deformed state and the tangential traction in the cohesive zone element. 14. T = exp − φ= ∆=0 T (∆) d∆ φ δ exp(1) . only one constitutive relation between the traction T and the opening displacement ∆ is used.02 0. 14.max Tt /Tt. The workofseparation φ is the dissipated energy after complete opening.01 0 −0. resulting in diﬀerent normal and tangential openings for diﬀerent local bases.04 3 4 3 4 3 4 1 2 1 2 1 2 0.2 0.25 0.16 : Inﬂuence of choice of reference plane.15 0.1 0.1 0. φ δ ∆ δ ∆ δ ∆ = ∆e ∞ .25 0.04 −0.15 0. the bottom plane (line between node 1 and node 2) and the upper plane (line between node 4 and node 5).02 −0.17 : Cohesive zone law for large deformations.max 0.05 0 0 ip1 ip2 0. 14.4 0. A large displacement formulation is therefore proposed to resolve the ambiguity induced by the choice of a local basis.05 Tt /Tt.max 0.2 0.3 0. A unit vector with e is deﬁned between two associated material points at the interface.3 0. The cohesive zone law is based on the normal traction relation of the exponential cohesive zone law discussed before.05 0.4 5 5 5 0.02 0 0. The choice of the local basis has a pronounced inﬂuence on the decomposition of the opening displacements.
i. In a twodimensional cohesive zone.18 : Large deformation cohesive zone response for elastic and history dependent behavior. Intermediate values of d represent a mixedmode opening. Fig. The tractionopening relation is extended with this modemixity d. Fig. the cohesive zone law may show an irreversible response. 14. Both cases are shown in the ﬁgure below. Parameter d has a value between 0 (modeI) and 2 (modeII). i. Modemixity The opening mode of the cohesive zone is quantiﬁed by a modemixity parameter d. where a parameter α controls the inﬂuence of modemixity behavior: the interface is stronger (α > 0) or weaker (α < 0) in modeII than in modeI. plastically damaged (ii). 14. elasticitybased damage (i) and unloading with the initial stiﬀness of the cohesive zone. d1 and d2 are the components of the normals n1 and n2 of the two cohesive zone edges perpendicular to ∆.165 Upon unloading.e.19 : Modemixity parameter for a twodimensional cohesive zone.e. d = d1 − d2  φ ∆ ∆ T = exp − δ δ δ exp α d 2 . Two types of irreversible behavior are considered: linear elastic unloading to the origin.
Most applications comprise small deformations and quasistatic loading. This phenomenon is also studied with cohesive zone models. After the forming steps.5 Applications Cohesive zones are used to study deadhesion between a wide range of materials under a large scope of loading conditions. leading to the loss of protective and attractive properties of the product which is unacceptable. where the total dissipated energy (area under the curve) depends on the parameter α. the total dissipated energy depends on α. Obviously. the traction is plotted for three values of α in modeII. 14.5. Large deformations occur in the next two examples : the delamination of the polymer coating on a steel sustrate. In the case of mixedmode or modeII (β = 0o ) opening of the cohesive zone.0 α = −0. as is used in beverage cans.2 α = 0. the repeated switching causes cyclic thermal loading and leads to deadhesion between the solder and the connection pad.0 α = −0.1 Polymer coated steel for packaging Traditionally. the opening of a foodcan. the product is cleaned and lacquered to prevent corrosion. beverage and food cans. In electronic devises. The results are shown in the ﬁgure below. delamination may occur. The value of the potential was taken to be φ = 100 Jm−2 . making subsequent lacquering superﬂuous. In the right ﬁgure. and. Repeated cyclic loading with small deformations can lead to fatigue damage and deadhesion between two entities. 14.166 To investigate the modemixity behavior. combined with a damage evolution law. 14.2 T [MPa] 50 40 30 20 α = 0. products like aerosols. where the lid is opened by peeling oﬀ a sealing foil. If delamination .20 : Inﬂuence of modemixity on cohesive zone response. Because the coating is subjected to the same deformation processes as the metal substrate. beer caps and luxury products are made from sheet metal. more and more products are made from polymer coated sheet metal. Currently. This implies considerable cost savings and eliminates the emission of volatile organic compounds. The two faces are kept parallel to each other and β is deﬁned as the angle between one of the faces and the vector ∆. the energy dissipated in modeI (β = 90o ) is independent of α and equals φ.2 α = 0. to give its surface a glossy appearance and to assure good printability.2 80 0 30 β [o] 60 90 0 0 2 4 ∆ [µm] 6 8 10 Fig. 120 110 W [Jm−2] 100 90 10 α = 0. a single cohesive zone is opened under an angle β and the total dissipated energy is quantiﬁed.
14. Fig. the processing routes. The protact ring is made of polymer coated steel with a PolyPropylene (PP) layer at the outside and a PolyEthyleneTerephthalate (PET) layer at the inside of the can. 14. The EPOL closure must be peelable with a force not exceeding 25 N under an angle of 135o .23 : Easy Peeloﬀ Lid in closed and opened state. Fig. which also is a common requirement in the industry. The EPOL is easy to open and to produce and consists of two parts: the protact ring and the aluﬁx.167 can be predicted. This research is described in detail in the MScthesis of S. van den Bosch [8] and papers. 14.6 bar at sterilization temperature. The thermoplastic PP layer of the protact ring and aluﬁx foil are heat sealed together.2 Easy peeloﬀ lid Many food containers are sealed with a socalled Easy PeelOﬀ Lid (EPOL).22 : Deepdraw simulation and delamination in polymersteel interface. achieving an airtight closure. The aluﬁx is a membrane. 14. Bijker [4]. This research is described in detail in the PhDthesis of M.42. Fig.5. parameters and tooling can be adjusted to prevent it. . which can be found in the References.J. The EPOL closure must also be airtight and resist a pressure buildup in the can of about 2.21 : Food and beaverage cans. which is the industrial standard for opening peelable food containers. consisting of aluminum which is coated with a peelable PP layer at one side of the foil.
Cyclic loading leads to fatigue damage and ﬁnally to failure. the fatigue damage is assessed with a cohesive zone approach. • Bending of the board induces shear and tensile stresses in the solder joints.E.5. . internal stresses build up in the solder. Erin¸ [23] and related c papers. where thermal loading of the system leads to deformation and results in plastic strains. Instead of using phenomenological laws like MansonCoﬃn. miniaturization is the major driving force in consumer electronics design and production.3 Solder joint fatigue Solder joints provide electrical. 14. thermal and mechanical continuity in electronic packages. This research is described in detail in the MScthesis of M. where solder balls are employed. • Cyclic thermomechanical loading evokes creepfatigue damage or creep rupture. Eﬀorts in decreasing component dimensions have led to the development of ball grid array (BGA) and ﬂip chip packages. These solder balls are subjected to diﬀerent types of loading: • Thermal cycling due to repeated power switching evokes heat related phenomena: the mismatch in the coeﬃcient of thermal expansion between the package components causes cyclic mechanical strains.24 : Ball grid array (BGA) and FEmodel of one solder ball with cohesive zones.168 14. which can be found in the References. The ﬁgure below shows a typical bump between two pads. Fig. • As a result of the multiphase nature of Sn based solder alloys and the thermal anisotropy of βSn. Today.
PhD Thesis.L. pp 12201234. pp 171180. of Solids and Structures. G. M. P. pp 302314. Vol 40. 1986..J.G.J.A cohesive zone model with a large displacement formulation accounting for interfacial ﬁbrillation. Clark.Bibliography [1] Barenblatt. P. R. [8] Bosch van den. Computational modelling of impact damage in brittle materials.. Advances in Applied Mechanics. M. R. Schreurs. [2] Barsoum. Vol 33. G. pp 380388. [6] Bosch van den. 1976. M...Interfacial characterization of prestrained polymer coated steel by a numericalexperimental approach.11.G. [10] Broek. Geers. Fractography and fracture mechanics.S. [7] Bosch van den.Interfacial delamination in polymer coated metal sheet. Vol 73. D. Int.G. M. Geers. Elementary engineering fracture mechanics.G.A. [3] Bates. J. Comp. Vol 42. Trans. M. 169 . Numerical Methods in Engineering.L. 2007.. Kluwer Academic Publishers.D.. Van Maris.J. An Introduction to Fatigue in Metals and Composites.F. P.G. 2008.J. [12] Camacho. The practical use of Fracture Mechanics. M. The mathematical theory of equilibrium cracks in brittle fracture.On the development of a 3D cohesive zone element in the presence of large deformations.J.Mech. Engineering Fracture Mechanics. Martinus Nijhoﬀ Publishers. 4thed. Kardomateas G. pp 119... 1996. 1989. Int.J.D. [9] Bosch van den.Mech. M. 2007. Mech. S.G.P. M. M. Eindhoven University of Technology. W.I.G.D. M.An improved description of the exponential Xu and Needleman cohesive zone law for mixedmode decohesion. Vol 10. [5] Bosch van den.G. 2008. Geers. Schreurs..G. [13] Carlson. 1996... 1969. Eur..T. Chapman and Hall. MT 08. Schreurs. ATM. Geers. 2008. D. Vol 7. MSc Thesis. R.J.J. [11] Broek. On the use of isoparametric Finite Elements in Linear Fracture Mechanics.H.C. Vol 62. pp 28992938... J. P.Study of the opening behavior of food containers with Easy PeelOﬀ Lid. [4] Bijker.. M. pp 55129.J. Eindhoven University of Technology.J..D. Schreurs. 2006. Vol 26. Ortiz. 1962.Mat.
J. M. 2004. of Basic Engineering.E. John Wiley and Sons. W. Yielding of steel sheets containing slit. [28] Griﬃth..D. E. P. G.Microstructural damc age analysis of SnAgCu solder joints and an assessment on indentation procedures.A. [29] Hertzberg. J. of Fracture Mechanics. PhD Thec sis.C. Solids and Structures. Clark H. M. Vol 221. Schreurs..J.J.170 [14] Donahue. Eindhoven University of Technology. Int. Structures.Integrated numericalexperimental analysis c of interfacial fatigue fracture in SnAgCu solder joints.M. Kluwer Academic Publishers. ASTM STP 486.Q. Deformation and fracture mechanics of engineering materials. 1976. M. Zhang.Sc. On the crack extension in plates under plane loading and transverse shear. Int. M. Basic Eng. Vol 89.D. Schreurs. Vol 44. Zhang. . an introduction. Phil. Ratwani M. J.E. [19] Erdogan. Kearney V.. M. [22] Erin¸. 4th ed. Sih.J.Electr. [30] Hertzberg. pp 100108. 1993. P. McEvily A. John Wiley and Sons.G. [16] Elber. Int. 1978. Microelectronics Reliability. Vol 16. [24] Forman. J. pp 230242. J. Kumble R. M.E.Characterization and fatigue c damage simulation in SAC solder joints.... Vol 6. G.G. 1960. pp 209219. Roy.. M. ASME. [25] Gdoutos.. 1970.. 1921.. [18] Erdogan. pp 56805694. 1971. 2005. Numerical analysis of crack propagation in cyclicloaded structures. [26] Gordon. G. F. pp 459464. of Fracture Mechanics. pp 693700. 2007. Fatigue crack closure under cyclic tension. Fatigue and fracture of cylindrical shells containing a circumferential crack. F..Q.E. Vol 8. Crack opening displacement and the rate of fatigue crack growth.G. Chichester. pp 163. pp 3745.D.E.. Vol 85. 2007.:Mat. Trans. 1996... R. Geers.G.. J.G. Penguin Books.Mat.G. [20] Erin¸. J. W. Richard W. Trans. D. Geers. 1967. Engineering Fracture Mechanics.J. pp 12871292. P.E. [23] Erin¸. Soc. 1976. [15] Dugdale.Thermomechanical fatigue failure of interfaces in leadfree solders. Penguin Books.S. Richard W. Fracture Mechanics . R. 1963. The signiﬁcance of fatigue crack closure. Vol 2. Engle R.J. 1970..E. [21] Erin¸. [27] Gordon.. Vol 44. Deformation and fracture mechanics of engineering materials.. Geers. 519527.. [17] Elber. pp.J.G.M. The new science of strong materials. Schreurs. J.. pp 379392. Atanmo P. Damage Tolerance in Aircraft Structures. of the Mechanics and Physics of Solids. A. 1972. Vol 8.
G.A. 2005. Vol 58. I. Y.J. Proc. Suppl. Geubelle.A. pp 190223. 1985.. pp 640657. J. J. [41] Knott. 1960.. S.Y. Vol 6. pp IV63. [45] McMaster (Ed. 1954.. [47] Miner. Fracture characteristics of three metals subjected to various strains.R. 1958. Cook.. M.W. 1958. New York. Choi. Stress intensity factors handbook. Kies. G. 2002. Vol 36.W. Journal of Applied Mechanics. strain rates. 1993. pp 3148. J. [36] Irwin. Smith. pp 34.R. et al. Pergamon. Advanced fracture mechanics. pp 5767. [44] Matous. Oxford University TTss. pp 551590. of fatigue. [32] Inglis. William H.171 [31] Hutchinson. PhD Thesis TU/e. A total dissipated energy theory of fatigue crack growth in ductile solids.H. [33] Irwin. Singular behavior at the end of a tensile crack in a hardening material.2. Popelar. N.F. 2003. [39] Kanninen.R. P. temperatures and pressures.. ..).109. of the Mechanics and Physics of Solids. Johnston. 1993(jul).L. M. pp 1331. K. [40] Klingbeil... [43] Lim. Continuous and discontinuous modelling of ductile fracture. 1968.. H.. Vol 21. Welding J. 1973.. 7th Sagamore Conf. J. Low cycle fatigue test for solders using noncontact digital image measurement system. Fundamentals of fracture mechanics. for Numerical Methods in Engineering. Res. pp A159A164. Application of singular quadratic distorted isoparametric elements in linear fracture mechanics.1 and Vol.. J. J. Vol 33.. [42] Kurz. Kies. G. Wolfgang Debonding along a ﬁber/matrix interface in a composite. International Journal for Numerical Methods in Engineering. Engineering Fracture Mechanics.. Transactions of the Institute of Naval Architects.W. [35] Irwin.L. N. [34] Irwin. Handbook of physics. WFW 93. Vol 55. pp 1935.K.F. Int. 1945. Int. 2006. Butterworths.E. [46] Mediavilla.A. C. Multiscale modeling of particle debonding in reinforced elastomers subjected to ﬁnite deformations. [38] Kanchanomai. pp 219241. Cumulative damage in fatigue. Vol 65. Stresses in a plate due to the presence of cracks and sharp corners. Fracture strength relative to onset and arrest of crack propagation. J.. Plastic zone near a crack and fracture toughness. G. C. Vol 24. Vol. C. 1985. Gordon R. 1987. Vol 16. Nondestructive Testing Handbook. Proceedings of the American Society for Testing Materials. [37] Johnson. pp 24732499... Vol 12.R. [48] Murakami. I. 1913.
pp 461465. [60] Sih.F. of Applied Mechanics. J. P.P. G. Vol 10. J. pp 305322. [59] Sih. Stress concentration design factors. The relation between crack growth resistance and fracture process parameters in elasticplastic solids. [65] Tvergaard. Pergamon.A. Vol 125. Hutchinson. of composite materials. [58] Schijve. J. 1963. Wiley. Engineering. J. Mech.. pp 230244. 1973. Engineering Fracture Mechanics. 1953. of Basic Engineering.E. Vol 7.. Handbook of stress intensity factors. Solids. V. J.. R. St. Del Research.W. Appl. Vol 16.R. [62] Tada. 2002. 2002. pp 544550. . Paris. J. 1973. ASME. pp 203213. [55] Rice.R.C. 2nd Ed. Paris Productions. 1985.. 1992. J. J.K. N. In: WCCM5. [53] Peterson. [56] Roberts. Vol 14. [66] Vinals.P. pp 112.R.. H. Vol 28. [52] Peterson. Y.. Fracture analysis of unidirectional composites. John Wiley. pp 13771397. [64] Tvergaard. J.172 [49] Murakami. R. Inc. P. The velocity of brittle fracture. [63] Tada.C. Vol 171. Compendium of stress intensity factors. A.J. Int. A path independent integral and the approximate analysis of strain concentrations by notches and cracks. Vol. Chen. Eﬀect of ﬁbre debonding in a whiskerreinforced metal. Some formulas for the crack opening stress level.R. ASME Trans. of the Mechanics and Physics of Solids. 1981. Vol 35. The stress analysis of cracks handbook. New York. [50] Neuber. [57] Rooke. Vol 40. Stress concentration factors. D. pp 820821. 1974.3. The stress analysis of cracks handbook. D. Strain energy density factor applied to mixed mode crack problems. [51] Paris. 1974..E. Bethlehem. Paris. 1968. G. E. A critical analysis of crack propagation laws. Theory of stress concentrations for shearstrained prismatical bodies with arbitrary nonlinear stressstrain law. London. Her Majesty’s Stationary Oﬃce.. G. Stress intensity factors handbook. 1968. 1976. Rosengren. H. Trans. 1954. Mech. Lehig University. G. [61] Sih.. J. G. Pennsylvania 18055.C. F. of Fracture. V.. Irwin. High performance fracture approach to fatigue crack analysis and life prediction. G. Louis. Cartwright. P. 1973. Erdogan.Y. H. J. pp 528534. J. Materials Science and Engineering A. 1990. Phys. D.. Wells. 1992. 1961. pp 379386. Vol 85.. Irwin..C. Plane strain deformation near a crack tip in a powerlaw hardening material. [54] Rice.C.
A.173 [67] Wasiluk.. British Welding Research Ass. Philosophical Magazine. [69] Xu. 1993. T. . Model Simul Mater Sci Engng. 1951.H.. The fracture process in elasticplastic materials under biaxial cyclic loading. A. Void nucleation by inclusions debonding in a crystal matrix.. pp 739750. [70] Yoﬀe. B. The moving Griﬃth crack. Vol 42. pp 111132. X. 1963. Application of fracture mechanics at and beyond general yielding. VOL 1. [68] Wells. Rep. MI3.P. 2003. Needleman.. Hoshide. A. E.
: : : . and interpreting fracture surface topography” Copied from Internet : www.17 12.3 9.3 2.23 139 139 : : : : : : : .7 12.de (2009) Copied from Internet : Univ. of Strathclyde (2006) Copied from Kanninen [39] Copied from Hertzberg [30].10 1.1 12.External sources of ﬁgures and other data Source ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg 1.12 2. Kanninen [39] Copied from Kurz [42] Copied from Mediavilla [46] Copied from Hertzberg [30] Unknown source Unknown source Unknown source Data taken from Hertzberg [29] Figures copied from Hertzberg [29] ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg ﬁg page page 3.1 3.7 1.2 8.16 11.5 12. Gdoutos [25]. .ventioelde. measuring.8 1.1 2.5 11.4 : : : : : : : Copied from Internet Copied from Internet (MaTe website) Copied from Kanninen [39] Unknown source Copied from Hertzberg [30] Copied from Hertzberg [30] Copied from Internet : Google Books : Derek Hull ”Fractography: observing.
.
r. After using tha CauchyRiemann equations and a second derivation w.t. The same can be deribed for the imaginary part of f . x1 and x2 are calculated ﬁrst. ∂φ ∂ζ +i ∂x2 ∂x2 → f ′ (z) = ∂f ∂f = −i ∂x1 ∂x2 → ∂φ ∂ζ −i ∂x2 ∂x2 ∂ζ ∂φ =− ∂x1 ∂x2 → ∂φ ∂ζ =− ∂x2 ∂x1 → → ∂2φ ∂2ζ 2 = − ∂x ∂x ∂x2 1 2 a1 .r.t.Appendix A Laplace equation The derivatives of the complex function f w. = ∂f ∂f ∂z ∂f = = = f ′ (z) ∂x1 ∂z ∂x1 ∂z ∂f ∂z ∂f ∂f = =i = if ′ (z) ∂x2 ∂z ∂x2 ∂z ∂ζ ∂φ +i = −i ∂x1 ∂x1 CauchyRiemann equations ∂φ ∂ζ = ∂x1 ∂x2 ∂2φ ∂2ζ 2 = ∂x ∂x ∂x1 1 2 ∂2φ ∂2φ + 2 =0 ∂x2 ∂x2 1 and . x1 and x2 . it follows that the real part of the complex function f satisﬁes the Laplace equations.
a2 .
22 = . ∂ψ ∂z ¯ ∂ψ ∂ z ∂ψ ∂ψ + = + ∂z ∂x1 ∂ z ∂x1 ¯ ∂z ∂z ¯ ∂ψ ∂ψ ∂z ∂ ∂ψ ∂ψ ∂ z ¯ ∂2ψ ∂2ψ ∂2ψ ∂ + + + =2 + + ∂z ∂z ∂ z ∂x1 ∂ z ∂z ¯ ¯ ∂ z ∂x1 ¯ ∂z∂ z ¯ ∂z 2 ∂z2 ¯ 2ψ 2ψ ∂ ∂ψ ∂ψ ∂z ∂ ∂ ∂ψ ∂ψ ∂ z ∂ ¯ + + + =i − 2 ∂z ∂z ∂ z ∂x2 ∂ z ∂z ¯ ¯ ∂ z ∂x2 ¯ ∂z ∂z2 ¯ ¯ ∂ψ ∂ z ∂ψ ∂ψ ∂ψ ∂z i− i + = ∂z ∂x2 ∂ z ∂x2 ¯ ∂z ∂z ¯ ∂ ∂ψ ∂z ∂z ¯ ∂ψ ∂ ∂ψ ∂ψ ∂2ψ ∂2ψ ∂2ψ i− i + i− i =2 − − 2 ∂z ∂z ∂z ¯ ∂x2 ∂ z ∂z ¯ ∂z ¯ ∂x2 ∂z∂ z ¯ ∂z 2 ∂z ¯ a3 ψ.Appendix B Derivatives of Airy function The general solution of the complex biharmonic equation is an expression for the Airy function in terms of complex functions. ¯ ∂ψ 1 ¯ ¯ = 2 z Ω ′ + Ω + ω′ ∂z ∂ψ 1 ¯ ¯ = 2 Ω + zΩ ′ + ω′ ∂z ¯ ∂2ψ ¯ = 1 Ω′ + Ω′ 2 ∂z∂ z ¯ ∂2ψ 1 ¯ = 2 z Ω ′′ + ω ′′ ∂z 2 ∂2ψ 1 ¯ ¯ = 2 z Ω ′′ + ω ′′ ∂z2 ¯ The derivatives w.t. ψ= 1 2 ¯ zΩ + zΩ + ω + ω ¯ ¯ The derivatives w.r.2 = ψ.12 = ψ.t. the coordinates x1 and x2 can now be calculated.1 = ψ.11 = ψ. the complex variables z and z can be determined.r.
a4 .
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.