STRESS CONCENTRATION

- Cracks always initiate at points of stress concentration.

- A crack, once initiated, becomes an intense stress

concentrator itself.

- Two fundamental cases of plane elasticity:

INFINITE PLATE CONTAINING A CIRCULAR HOLE

(Kirsh, G, (1898), V.D.I., 42, 797-807)

→ Stress Concentration

INFINITE PLATE CONTAINING AN ELLIPTICAL HOLE

(Kolosoff, G.V., On an application of complex function theory to a plane
problem of the mathematical theory of elasticity, Yuriev, 1909;
Inglis, C.E., (1913), Stresses in a plate due to the presence of cracks and
sharp corners,Transactions of the Royal Institute of Naval Architectes, 60,
219-241)

→ Stress Concentration

concentration 1
- In the limit of (minor axis) / (major axis) → 0 :

INFINITE PLATE CONTAINING A CRACK

(Wieghardt, K., (1907), Z. Mathematik und Physik, 55, 60-103; translated by

Rossmanith (1995), On splitting and cracking of elastic bodies, Fatigue Fract.
Engrg. Mater. Struct., 18, 1371-1405)
Muskhelishvili, N.I., Some basic problems of the mathematical theory of
elasticity, in Russian 1933 (in English 1953, Noorhoff-Groningen).
Westergaard, Bearing Pressures and cracks, (1937), J. Applied Mechanics,
6, A49-53.
Williams, M.L. (1952), Stress singularities resulting from various boundary
conditions in angular corners of plates in extension, J. Applied Mech., 19,
526-528.
Williams, M.L. (1957), On the stress distribution at the base of a stationary
crack, J. Applied Mech., 24, 109-114.)

→ Stress Intensification

concentration 2
 INFINITE PLATE CONTAINING A CIRCULAR HOLE

(Kirsh, G, (1898), V.D.I., 42, 797-807)

- Consider infinite plate containing a circular hole of radius R

and subject to a remote tensile stress σx = σ.

- Consider portion of plate within concentric circle of radius

R' >> R so that stress field is not perturbed by hole
(Saint- Venant’s Principle)

- Stress field at r = R' (Mohr's circle):

σ
σr = (1 + cos2θ )
2
σ
σ rθ = − sin2θ
2

concentration 3
- Decompose problem into:

Problem (1) Problem (2)

σ r = σ / 2, σ rθ = 0 σ σ
σr = (cos2θ ), σ rθ =− sin2θ
2 2

- Solution problem (1):

thick cylindrical tube under tension (Lamè):

σ R 2R'2 1 σ R'2
σr = − 2 2 2
+
2 R' −R r 2 R'2 −R 2
σ R 2R'2 1 σ R'2
σθ = + +
2 R'2 −R 2 r 2 2 R'2 −R 2

- For R' → ∞
σ R 2 
σ r = + 1 − 2 
2 r 
R 2 
σ 
σ θ = + 1 + 2 
2 r 

concentration 4
- Solution problem (2):

Assume Airy stress function:

Φ = f (r )cos2θ

Impose compatibility, ∇4Φ = 0,

2
 ∂2 1 ∂ 1 ∂ 2 
 + +  Φ=0
 ∂r 2 r ∂r r 2 ∂θ 2 
 

Substitute stress function and get the ordinary differential

equation to determine f(r):

2
 d2 1 d 4 
cos2θ  2 + − 2 f =0
 dr r dr r 


General solution:

C
f = Ar 2 + Br 4 + 2
+D
r

Boundary conditions:
σ
σ r (R' ) = + cos2θ
2
σ
σ rθ (R' ) = − sin2θ
2
σ r (R) = 0
σ rθ (R) = 0

concentration 5
Recall:

1 ∂Φ 1 ∂ 2Φ
σr = + +
r ∂r r 2 ∂θ 2
∂ 2Φ
σθ = +
∂r 2
1 ∂Φ 1 ∂ 2Φ
σ rθ = + 2 −
r ∂θ r ∂r∂θ

Substitute into general solution to get constants:

σ σ σ
For R' → ∞ : A = − , B = 0, C = − R 4 , D = R 2
4 4 2

- Solution problems (1) + (2):

σ  R 2  σ  R4 R 2 
σ r = 1 − 2  + 1 + 3 4 − 4 2 cos2θ
2 r  2 r r 
σ  R 2  σ  R 4 
σ θ = 1 + 2  − 1 + 3 4 cos2θ
2 r  2 r 
σ  R4 R 2 
σ rθ = − 1 − 3 4 + 2 2 sin2θ
2 r r 

concentration 6
For r = R:

σr = 0
σ θ = σ (1 − 2cos2θ )
σ rθ = 0

Maximum stress: σ θ = 3σ for θ = π/2, 3π/2

Minimum stress: σ θ = −σ for θ = 0, π

Note:

- stress concentration factor = 3, independent of R

- compression for −π/6 ≤ θ ≤ π/6 and −5π/6 ≤ θ ≤ 7π/6

- circumferential stress for θ = π/2:

σ  R2 R 4 
σθ = 2+ +3 4
2  r2 r 

⇒ hole has a localized character

for r = 4R: σ θ = σ (1+ 0.04)

⇒ solution applicable to finite plates with width > 4R.

- stress field satisfies plane-strain and generalized plane

stress.

concentration 7
- INFINITE PLATE WITH CIRCULAR HOLE SUBJECT TO
BIAXIAL STRESS

Apply superposition principle to get stresses at r = R:

1) Biaxial tension:

Uniform stress: σ θ = 2σ

2) Tension/compression (pure shear):

Maximum stress: σ θ = 3σ + σ = 4σ
for θ = π/2, 3π/2

Minimum stress: σ θ = −σ − 3σ = −4σ

for θ = 0, π

concentration 8
 INFINITE PLATE CONTAINING AN ELLIPTICAL HOLE

(Kolosoff, G.V., On an application of complex function theory to a plane

problem of the mathematical theory of elasticity, Yuriev, 1909;
Inglis, C.E., (1913), Stresses in a plate due to the presence of cracks and
sharp corners,Transactions of the Royal Institute of Naval Architectes, 60,
219-241)

Hp: elliptical hole with: a = major axis, b = minor axis.

- Solution:
Kolosof (complex function theory);
Inglis (Conformal Mapping, elliptical coordinates)
(see Carpinteri, Meccanica dei materiali e della frattura, 1992, Pitagora,
for details).

- For β = π/2 (tensile stress perpendicular to major axis):

σ
 1- m2 - 2m + 2cos2θ 
σ θ =  2
σ

θ  1 + m - 2mcos2θ 

m=(a-b)/(a+b).
a = major axis, b = minor axis

concentration 9
Maximum circumferential stress:

 1+ m   a
maxσ θ = σ θ (0) = 1 + 2 σ = 1 + 2 σ
 1- m   b

- Other cases:

Tensile stress  a
perpendicular to σ max = σ 1 + 2 
major axis  b
Tensile stress  b
perpendicular to σ max = σ 1 + 2 
minor axis  a

a
Uniform stress σ max = 2 σ
b

 a
Pure shear σ max = 2σ1+ 
 b

(a + b)2
Pure shear parallel to σ max =σ
major and minor axes ab

- For a = b, solution for a circular hole (Kirsh);

- For b/a → 0 and tensile stress perpendicular to major

axis:
stress intensification;

- For b/a → 0 and tensile stress perpendicular to minor axis:

uniform stress equal to applied load,

no stress concentration

concentration 10
 STRESS INTENSIFICATION

Plate with a crack: stress intensification at the crack tip

- What is the critical load in the cracked plate?

- What is the "fine structure" of the stress field at the crack

tip?

- What is the power of the singularity?

(e.g., 2D problem of a point load acting on a semi-infinite plane (Kelvin):
stresses = O(r −1), displacements = O(log r))

intensification 1
Traditional design approach:
(2 parameters: σapplied, σu)

σapplied < σu

σu = yield or tensile strength

⇒ (a): σapplied < σu / 3 → safe

⇒ (b): σapplied < σu / (1+a/b) → safe
⇒ (c): σapplied ≠ 0 → unsafe for any σy

???

Fracture Mechanics approaches in design:

(3 parameters: σapplied, GIC (or KIC), a)

- Energy Approach:

the crack propagates when the energy available

for crack growth
overcomes the material resistance (fracture toughness)

- Stress Intensity Factor Approach:

the crack propagates when a local measure

of the singular stress field
reaches a critical value (fracture toughness)

intensification 2
- Energy Approach:
(Griffith, A.A., The phenomena of rupture and flow in solids, Philosofical
Transactions, Series A, vol. 221, 1921, 163-198;
Griffith, A.A., The theory of rupture, First Int. Congress of Applied
Mechanics, Delft, 1924, 55-63;)

- Stress Intensity Factor Approach:

(Wieghardt, K., (1907), Z. Mathematik und Physik, 55, 60-103; translated by
Rossmanith (1995), On splitting and cracking of elastic bodies, Fatigue Fract.
Engrg. Mater. Struct., 18, 1371-1405)
Muskhelishvili, N.I., Some basic problems of the mathematical theory of
elasticity, in Russian 1933 (in English 1953, Noorhoff-Groningen).
Westergaard, Bearing Pressures and cracks, (1937), J. Applied Mechanics,
6, A49-53.
Williams, M.L. (1952), Stress singularities resulting from various boundary
conditions in angular corners of plates in extension, J. Applied Mech., 19,
526-528.
Williams, M.L. (1957), On the stress distribution at the base of a stationary
crack, J. Applied Mech., 24, 109-114.)

intensification 3
CRACK TIP SINGULARITY IN AN INFINITE PLATE
CONTAINING A CRACK
WESTERGAARD METHOD
(Westergaard, Bearing Pressures and cracks, (1937), J. Applied Mechanics,
6, A49-53.)
(see Carpinteri, Meccanica dei materiali e della frattura, for details)

Hp:
homogeneous, isotropic, linearly elastic body
plane stress
crack length = 2a
biaxial load:
σ y (z = ∞ ) = σ
σ x (z = ∞ ) = kσ
with k = real constant, - ∞ ≤ k ≤ ∞

Problem: define stress field at crack tip

intensification 4
Stress field ahead of the crack tip in an infinite cracked
plate (symmetry about x axis)

σx =
KI
2π r
( )
cos θ2 1- sin θ2 sin 32θ + 2B

σy =
KI
2π r
(
cos θ2 1 + sin θ2 sin 32θ)
KI
σ xy = sin θ2 cos θ2 sin 32θ
2π r

Note:

- General expression for the stress components:

KI
σ ij = fIij (θ )
2π r

- inverse square root singularity in all stress components:

- power of the singularity and fIij(θ) depending on crack face

boundary conditions and unaffected by remote boundary
conditions

- stress field univocally defined by KI (now still unknown):

→ KI is a measure of the singular stress field

→ single parameter description of crack tip conditions

- KI units: [F] [L] -3/2

- Singularity dominated zone

intensification 11
Recall g(a) a = KI π then:

K I = σ πa

KI = Stress Intensity Factor (SIF) in an infinite cracked plate

subject to biaxial loading (symmetry about x axis):

Note:

- σx(∞) does not affect KI.

- KI and stress field at the crack tip depend on crack length a

- σx = σ(k-1) for θ = π, along crack faces:

σx = -σ uniaxial load (k = 0)

σx = 0 biaxial load, (k = 1 or σx(∞)=σy(∞)=σ)

intensification 13
 EFFECT OF FINITE SIZE:
MODE I PROBLEMS IN FINITE BODIES

- Stresses at the crack tip:

KI
σ ij = fIij (θ )
2π r

- Stress Intensity Factor in finite body:

K I = σ πa F

F dimensionless function, typically a polynomial expression,

depending on geometry and loading conditions.
intensification 14
Examples:

- Cracked strip in tension:

1/2
 πa 
K I = σ πa  sec 
 2h 
1/2
a  πa 
F = F  =  sec  guess based on
h  2h  numerical solution

a
F 
h

- Three point bending beam:

Pl a
KI = F 
th3/2  h 
1/2 3/2
a a a
F  = 2.9  - 4.6   +
h h h
5/2 7/2 9/2
a a a
+ 21.8   - 37.6   - 38.7  
h h h

intensification 15
 CRACK OPENING DISPLACEMENT

- From plane stress elasticity:

∂u y 1
εy = = (σ y −νσ x )
∂y E
then:

u y = ∫ ε y dy + rigid body motion terms

- substituting the stresses in the singularity dominated

region:

1 ν
u y = ∫ ε y dy = ∫ (ReZ I + yImZ '
I − B ) dy - ∫ (ReZ I - yImZ '
I + B) dy
E E

- It is easy to check that the following uy satisfies previous

equation:

2 1+ ν 1+ ν
uy = Im ZI − yReZI − By
E E E

intensification 16
- using the complex potential:

KI
ZI = 2ξ1/2 + Bξ + C
2π

- and polar coordinates:

2K I 1/2 θ θ
ZI = r (cos + i sin ) + Br(cosθ + i sin θ ) + C
2π 2 2

-so that:

1/2
2 K I 1/2
COD = u y (θ = π ) − u y (θ = −π ) = 4  r
π  E

Note:

- COD depends on KI/E and increases on increasing KI/E

- COD is parabolic along the crack, COD ≈ r1/2

- COD has a vertical tangent in r = 0

intensification 17
 ELEMENTARY MODES OF LOADING
APPLIED TO A CRACK

Mode I: Opening
symmetric loading about crack plane

Mode II: In-plane shear or in-plane sliding

anti-symmetric (or skew-symmetric) loading about x axis

Mode III: Tearing or out of plane shear

anti-symmetric (or antiplane) loading about x-z plane

intensification 18
 INFINITE CRACKED PLATE IN MODE II
(Westergaard, Bearing Pressures and cracks, (1937), J. Applied Mechanics,
6, A49-53.)

- Airy stress function: ΦII = − yReZII

- ………
- Stresses at crack tip:

σx = −
K II
2π r
(
sin θ2 2 + cos θ2 cos 32θ )
K II
σy = cos θ2 sin θ2 cos 32θ
2π r
σ xy =
K II
2π r
( )
cos θ2 1 - sin θ2 sin 32θ

K II
- σ ij = fIIij (θ )
2π r
- Stress Intensity Factor:

K II = τ πa
intensification 19
 STRESS FIELD AHEAD OF THE CRACK TIP

Mode I:
σx =
KI
2π r
(
cos θ2 1- sin θ2 sin 32θ + 2B )
σy =
KI
2π r
(
cos θ2 1 + sin θ2 sin 32θ )
KI
σ xy = sin θ2 cos θ2 sin 32θ
2π r
σ xz = σ yz = 0
σ z = ν (σ x + σ y ) in plane strain, σ z = 0 in plane stress

Mode II:
σx = −
K II
2π r
(
sin θ2 2 + cos θ2 cos 32θ )
K II
σy = cos θ2 sin θ2 cos 32θ
2π r
σ xy =
K II
2π r
(
cos θ2 1 - sin θ2 sin 32θ )
σ xz = σ yz = 0
σ z = ν (σ x + σ y ) in plane strain, σ z = 0 in plane stress

Mode III:
K III
σ xz = − sin θ2
2π r
K
σ yz = III cos θ2
2π r
σ x = σ y = σ xy = σ x = 0

intensification 20
 CRACK TIP DISPLACEMENT FIELDS

Mode I:
KI r θ  2θ
uy = sin  (κ + 1) - 2cos
2G 2π 2 2 
KI r θ  θ
ux = cos (κ - 1) + 2sin2 
2G 2π 2 2
uz = 0

Mode II:
K II r θ  θ
uy = − cos (κ − 1) - 2 sin2
2G 2π 2  2 
K r θ  θ
u x = + II sin (κ + 1) + 2 cos 2
2G 2π 2 2 
uz = 0
Mode III:
KIII 2r θ
uz = − sin
G π 2
ux = 0
uy = 0

where:

κ = 3−4ν (plane strain)

κ = (3−ν)/(1+ν) (plane stress)

intensification 21
 SIF EVALUATION

- Close form solutions:

complex function theory (conformal mapping, boundary

collocation method, Laurent series expansion,…

integral transforms (Fourier, Mellin, Hanckel transforms)

eigenfunction expansion

limited to very simple cases

- Computational solutions (FEM, BEM, FDM, …)

- Experimental solutions (photoelasticity, moire

interferometry,…)

- Fracture handbooks

Tada, Paris and Irwin, (1985), The stress analysis of cracks

handbook, Paris Prod., Inc., St. Luis (II edition).

Murakami (1987), Stress Intensity Factor handbook, Pergamon

Press, NY.

Rooke and Cartwright, (1976), Compendium of Stress Intensity

Factors, Her Majesty’s Stationary Office, London.

intensification 23
Tada, Paris and Irwin, (1985), The stress analysis of cracks
handbook

intensification 24
 STRESS INTENSITY FACTOR
SUPERPOSITION PRINCIPLE

- For linear elastic materials individual components of

stress, strain and displacement are additive.

- Similarly, stress intensity factors are additive:

K I = K (A) + K (B) + K (C) + ...

I I I

K II = K (A) + K (B) + K (C) + ...

II II II

K III = K (A) + K (B) + K (C) + ...

III III III
where:

K (i) , K (i) , K (i) (i = A, B, C, ….) are mode I, mode II and mode

I II III
III stress intensity factors for the ith applied load.

Example:

K I = K (SENT)
I

K I = K (SENB)
I

K I = K (SENT) + K (SENB)
I I

intensification 28
 SUPERPOSITION PRINCIPLE

Any loading configuration can be represented by

appropriate tractions applied directly to the crack faces

Proof:

intensification 31
 SUPERPOSITION PRINCIPLE

- Infinite plate subject to uniaxial load with a crack of length

2a oriented at an angle β with the x axis

- For β ≠ 0 the crack experiences Mode I and Mode II

- Introduce a new coordinate system x' - y' with x' coincident

with the crack orientation

- Define loads in the new system using Mohr's Circle

σ x ' = σ sin2 β
σ y ' = σ cos 2 β
σ x 'y ' = σ sinβ cosβ

- Stress intensity factors

K I = σ cos 2 β πa
KII = σ sinβcosβ πa
intensification 32
 MIXED MODE FRACTURE

Assumptions:
homogeneous, isotropic, linear elastic material
plane problem (Mode I + Mode II)

Griffith’s Energy Criterion applies only to collinear crack

growth

A Criterion for Mixed Mode fracture must define:

a) the direction of crack growth

b) the critical values of the fracture parameters

MAXIMUM CIRCUMFERENTIAL STRESS CRITERION

(Erdogan and Sih, (1963), On crack extension in plates under plane loading
and transverse shear, J. Basic Engineering, 85, 519-527)

mixed mode 1
 Example of Application

- Cracked plate subject to biaxial load σ1 - σ2

- Assume: m = σ1 / σ2
β = crack direction about σ2

KI = σ β π a
- Crack tip stress intensity factors:
K II = τ β π a

mixed mode 2
 Direction of crack growth

The crack propagates in a direction normal to the maximum

circumferential stress σθ

Special cases:

1) m = 1 , σ1 = σ2 ⇒ Mohr’s circle degenerates

into a point
symmetry about crack direction
σ β = σ1
⇒ KII = 0, θ=0
τβ = 0
⇒ collinear crack growth

2) m = 0, σ1 = 0 (uniaxial tension)
if β = 0 ⇒θ=0 collinear crack growth
lim
if β ≠ 0 ⇒ θ (m = 0, β ) = 70.6°
β → 0+
mixed mode 3
 Criterion for crack growth

The Mixed-Mode crack will propagate along θ when σθ =

σθcr(Mode I)

Fracture locus

- the fracture locus is symmetric about the K*I axis

- the fracture locus is not defined for KI < 0

mixed mode 4
 GRIFFITH ENERGY CRITERION
(Griffith, A.A., The phenomena of rupture and flow in solids, Philosofical
Transactions, Series A, vol. 221, 1921, 163-198;
Griffith, A.A., The theory of rupture, First Int. Congress of Applied
Mechanics, Delft, 1924, 55-63;)

- First law of Thermodynamics:

when a system goes from a nonequilibrium state to
equilibrium there will be a net decrease in energy

→ A crack can form in a body only if such a process causes

the total energy to decrease or remain constant (Griffith)

- Consider an infinite plate of unit width subject to biaxial

load in plane stress conditions.

Imagine to create a crack of length 2a while keeping the

remote displacements constant

The elastic strain energy released is proportional to the

strain energy contained in a circle of radius a (after Inglis):

2σ
2
We = πa
E

energy approach 2
The strain energy, W, of the cracked plate is then given by:

2σ
2
W = W0 - We = W0 - πa
E

where W0 is the strain energy of the uncracked plate.

- The energy required to create the crack surface is:

Ws = 4aγ s

- An incremental crack extension da is possible if:

dW dWs
+ ≤0
da da

σ2
2πa ≥ 4γ s
E

2γ sE
σ≥
πa

(substitute E with E’ for plane strain)

energy approach 3
Fracture Energy:

GIC = 2 γs

= energy required to create a unitary crack area.

(crack area = 2a; surface area = 4a)

- measure of the fracture toughness of the material

- to be defined through standard tests

Griffith equation:

Stress for equilibrium

crack growth:

GICE
σ max =
πa

GICE
σ> dynamic growth
πa
G E
σ < IC no growth
πa

energy approach 4
Stress for equilibrium
crack growth:

GICE
σ max =
πa

GICE
σ> dynamic growth
πa
G E
σ < IC no growth
πa

- crack growth is unstable: the load decreases on increasing

the crack length

- two asymptotes:

• for a →∞, σmax → 0

• for a →0 σmax → ∞ ????

analogy with buckling collapse vs. strength collapse:

for a<a0 the strength collapse precedes the fracture collapse

- The crack length a0 corresponding to σmax = σy ( = σp) is an

equivalent measure of the microcracks, flaws and defects of
the virgin material:

1 GICE
a0 =
π σ y2

energy approach 5
 MODIFIED GRIFFITH EQUATION

(Irwin, (1948), Fracture Dynamics, Fracturing of Metals, ASM Cleveland,

146-166;
Orowan,(1948), Fracture and strength of solids, Reports on Progress in
Physics, XII, p. 195.)

- In actual structural materials:

a) the energy needed to cause fracture is much higher than

the surface energy (orders of magnitude higher)

b) inelastic deformations arise around the crack front →

linear elastic medium with infinite stresses at the crack tip
unrealistic

- Griffith equation can be modified to include the plastic work

γP dissipated at the crack front

2E(γ s + γ p )
σ≥
πa

- Orowan estimated γP ≈ 103 γs in typical metals.

- The modified Griffith equation is correct only if:

the size of the plastic zone around the crack tip is very small
compared to the crack size (small-scale yielding conditions)

→ the details of the crack tip stress do not affect the stress
field in the elastic bulk of the medium
→ purely elastic solutions can be used to calculate the rate
of energy available for fracture (!!!!!!!!)

energy approach 6
 ENERGY BALANCE
IN BRITTLE FRACTURE

- Griffith criterion refers to a special case, i.e. infinite

cracked plate, biaxial loading, fixed-grip conditions

- In general, the First Law of Thermodynamics yields:

dE T dW dWs
= + ≤0
dA dA dA

A = crack area
ET = total energy
Ws = energy required to create new crack surfaces
W = total potential energy W=U-L
U = potential strain energy
L = potential of external forces

- From the definition of fracture energy:

dWs
= 2γ s = GIC
dA

- Energy criterion for brittle crack growth (Mode I):

dW
− ≥ GIC
dA

energy approach 7
 THE STRAIN ENERGY RELEASE RATE
(Irwin, (1957), Analysis of stresses and strains near the end of a crack
traversing a plate. ASME Journal of Applied Mechanics, 24, 361-364)

- Irwin introduced the Strain Energy Release Rate

dW
G =−
dA

as the energy available for an increment of crack extension,

given by the total potential energy released due to the
formation of a unit crack area

- G is also called Crack Driving Force or Crack Extension

Force

- Energy criterion for brittle crack growth (Mode I):

GI = GIC

energy approach 8
 ENERGY APPROACH VS. STRESS APPROACH
(CRITICAL CONDITIONS)

- The strain energy release rate is the fracture parameter

describing the global behavior of the body.

- The stress intensity factor if the fracture parameter

describing the local behavior (stresses, strains and
displacements near the crack tip) of the body

- Stress criterion for brittle crack growth (Mode I):

KI ≥ KIC

KIC = critical stress intensity factor

KIC is a measure of the material fracture toughness to be

defined through standard tests

- GIC and KIC are related through a fundamental relationship

2
KIC
GIC =
E

derived from:

GICE
σ≥ energy instability condition
πa
K
σ ≥ IC stress instability condition
πa
⇓
K IC = GICE

energy approach 9
 STRAIN ENERGY RELEASE RATE
IN A DCB SPECIMEN

- Assume built-in end conditions for the two arms.

From beam theory:

Pa3 4Pa3
Δ/2 = =
3EI EBh3

- The elastic compliance is:

Δ 2a3
C= =
P 3EI

- The strain energy release rate in the DCB specimen is:

P 2 dC P 2 6a 2 12 P 2 a 2
G= = =
2B da 2B 3EI B 2 Eh3

- Same conclusion from:

dW 1 d ⎛ 1 ⎞ 1 d ⎛⎜ P 2a 3 ⎟⎞ P 2 a 2 12 P 2 a 2
G =− =− ⎜ − PΔ ⎟ = = =
Bda B da ⎝ 2 ⎠ B da ⎜⎝ 3EI ⎟⎠ B EI B 2 Eh3

energy approach 17
- The critical load for crack propagation (Griffith criterion) is

1 GICEh3B2
Pcr =
a 12

⇒ unstable crack growth

Note:

Strain energy release rate and critical load have been

defined through a beam theory approximation.
If the geometry satisfies the assumptions of beam theory the
solution is correct.

However, the local fields at the crack tip are not correctly
given by beam theory

energy approach 18
 ENERGY RELEASE RATE IN A BODY SUBJECT
TO DIFFERENT LOADING SYSTEMS

Hp: self-similar crack growth

- From the stress intensity factors of a body subject to

different loading systems (A), (B), (C),…..

K I = K (A) + K (B) + K (C) + ...

I I I

K II = K (A) + K (B) + K (C) + ...

II II II

K III = K (A) + K (B) + K (C) + ...

III III III

and the relationship between G and K

K I2 K II2 K III
2
G= + +
E E 2G

⇒ the strain energy release rate is

(KI(A) + KI(B) + KI(C) + ....)2 (KII(A) + KII(B) + ...)2 (KIII

(A)
+ KIII
(B)
+ ....)2
G= + +
E E 2G

energy approach 19
- The strain energy release rate contributions for each mode
of fracture are addictive

K I2 K II2 K III
2
G = G I + G II + G II = + +
E E 2G

- The strain energy release rate contributions for each

loading system are not addictive, e.g. for mode I

GI(A)+(B)+(C)+... ≠ GI(A) + GI(B) + GI(C) + ...

energy approach 20
 NONLINEAR FRACTURE MECHANICS
THE COHESIVE CRACK MODEL
(Dugdale, D.S.: Yielding of steel sheets containing slits, J. Mechanics
Physics Solids 8 (1960), 100-104
Barenblatt, G.I.: The formation of equilibrium cracks during brittle fracture.
General ideas and hypotheses. Axially-symmetric cracks, J. Applied
Mathematics and Mechanics 23 (1959), 622-636.
Barenblatt, G.I.: The mathematical theory of equilibrium cracks in brittle
fracture, in H.L. Dryden and T. von Karman (eds.), Advanced in Applied
Mechanics, Academic Press, New York, 1962, pp. 55-129.)

ductile materials ⇒ Dugdale’s model

purely brittle materials ⇒ Barenblatt’s model

quasi brittle materials ⇒ Hillerborg’s Fictitious

(e.g. concrete) Crack Model

brittle matrix composites ⇒ Bridged crack model

quasi brittle matrix composites ⇒ Cohesive crack model

cohesive crack model 1

 Dugdale’s Model for ductile fracture

Fracture of elasto-plastic materials

Assumptions:

- Mode I crack in an infinite sheet under uniform tensile

stress
- the material is ductile and the plastic deformations
localized in a thin zone coplanar with the crack
- the plastic zone is modeled through a fictitious crack (of
unknown length) and a uniform distribution of cohesive
tractions σP = yield stress

Length of the cohesive zone

The length of the cohesive zone aP is calculated by

imposing the condition of smooth closure of the crack faces:

K I = K I σ + K Iσ P = 0

where: K Iσ = σ π (a + aP )

cohesive crack model 2

The SIF due to the plastic stresses is calculated using the
SIFs due to a pair of concentrated forces F acting at x:

1/2
F ⎛a+ x⎞
K I (A) = ⎜ ⎟
πa ⎝ a − x ⎠
1/2
F ⎛a−x⎞
K I (B) = ⎜ ⎟
πa ⎝ a + x ⎠

so that:
aP
σP ⎡ (a + aP ) + x (a + aP ) − x ⎤
K Iσ P = -
π (a + aP ) ∫ ⎢
(a + a ) − x
+
(a + a ) + x
⎥ dx
a ⎣ P P ⎦

(a + aP ) a
K Iσ P (a) = - 2σ P arccos
π (a + aP )

From the condition for smooth closure:

(a + aP ) a
σ π (a + aP ) − 2σ P arccos =0
π (a + aP )

a πσ
= cos
a + aP 2σ P

In the limit σ= 0 ⇒ aP = 0
σ → σP ⇒ aP → ∞
cohesive crack model 3
Performing a Taylor series expansion (if σ << σP):
2
a 1 ⎛ πσ ⎞
= 1− ⎜⎜ ⎟
a + aP 2 ⎝ 2σ P ⎟⎠

aP π 2σ 2
=
a + aP 8σ P 2

and, as a function of KIσ,

π K Iσ 2
aP =
8 σ P2

Length of the cohesive zone in critical conditions

Following Irwin's approach:

π K IC 2
aPC =
8 σ P2

where: K IC = GICE

δa

and: ∫ σ P d δ = σ Pδ a = GIC
0

with δa the crack tip opening displacement given by:

2
8σ a ⎡ ⎛ πσ ⎞⎤ KI
δ a= P ln⎢sec ⎜⎜ ⎟⎟⎥ or δ a= (if σ << σP)
πE ⎣ σ PE
⎝ 2σ P ⎠⎦

(use Castigliano's method)

cohesive crack model 4

 Barenblatt’s Model for purely brittle fracture
“In our work the question concerning equilibrium cracks forming
during brittle fracture of a material is presented as a problem in the
classical theory of elasticity, based on certain very general
hypotheses concerning the structure of a crack and the forces of
interaction between its opposite sides, and also on the hypothesis
of finite stresses at the ends of the crack, or, which amounts to
the same thing, the smoothness of the joining of opposite sides
of the crack at its ends” (Barenblatt, 1959)

Consider a Mode I crack in a homogeneous, isotropic,

linear-elastic infinite medium under uniform loading

Represent the atomic bonds holding together the two halves

of the body separated by the crack as cohesive forces
acting along the edge regions of the crack and attracting
one side of the crack to the other

b = interatomic distance (order = 10-7 mm)

E Eγ s
maximum intensity = ideal strength σ c ≈ ≈
π b
cohesive crack model 5
Determination of the cohesive forces

The accurate determination of the cohesive forces acting

along the edge regions is difficult

- Assumptions:

1) The dimension of the edge region is small in comparison

with the size of the whole crack

2) The displacements in the edge region, for a given

material under given conditions, is always the same for any
acting load

⇒ during crack propagation the edge region simply translate

forward

1) + 2) = small scale yielding assumption

3) The opposite sides of the crack are smoothly joined at the

ends or, which amounts to the same thing, the stress at the
end of a crack is finite

⇒ K = 0 (zero stress intensity factor)

These assumptions lead to the definition of the critical state

of mobile equilibrium which depends on the modulus of
cohesion depending on the integral of the cohesive tractions
along the edge region
cohesive crack model 6
Equivalence of Griffith and Barenblatt approaches

Willis (by means of the complex variable method) and

Rice (by means of the J integral) proved that:

Barenblatt’s theory based on atomic forces is equivalent to

Griffith’s energy approach provided the integral of the
cohesive forces is equal to the fracture energy
y*
∫ f dy = GIC
0

Barenblatt also showed that the cohesive forces essentially

have effect only on the displacement field close to the edge
of the crack and not on those in the main part of the crack

cohesive crack model 7

 Rice’s proof of the equivalence between Barenblatt’s
model and Griffith’s energy approach

- Assume the crack to be in a state of mobile equilibrium

- Consider cohesive forces σ(δ) acting along the crack
- Let δ∗ to be the separation distance beyond which the cohesive
tractions vanish

- Consider the path Γ for which:

dy = 0, ds = dx on Γ− and ds = -dx on Γ+ , T1 = 0, T2 = σ(δ) on Γ+ ,
T2 = -σ(δ) su Γ−

-Evaluate the J integral along Γ:

a +d a
⎛ ∂u ⎞ ∂u ∂u ∂u ∂u
J = ∫ ⎜ U d dy − Ti ⋅ i ds ⎟ = - ∫ T2 2 ds - ∫ T2 2 ds = - ∫ - σ (δ ) 2 (dx) - ∫ σ (δ ) 2 (-dx) =
Γ⎝
∂x ⎠ Γ− ∂x Γ+
∂x a
∂x d +a
∂x
a +d a +d
∂ ∂δ
=-∫ σ (δ ) (u 2 + - u 2 - )dx = - ∫ σ (δ ) dx ⇒
a
∂x a
∂x
δ*
J = ∫ σ (δ )dδ
0
⇒ The value of J which will cause crack extension is given by the
integral of the cohesive forces:

δ*
⇒ ∫ σ (δ ) dδ = GIC
0

cohesive crack model 8