“Measure what is measurable, and make measurable what is not so.

” - Galileo GALILEI

Linear Elastic Fracture Mechanics

Krishnaswamy Ravi-Chandar

Lecture presented at the

University of Pierre and Marie Curie
April 30, 2014

Center for Mechanics of Solids, Structures and Materials

Department of Aerospace Engineering and Engineering Mechanics
 The plan
• Introduction to fracture mechanics
– April 30, 2014
• Quasi-static instability problems (JB Leblond)
– May 6, 2014
• Introduction to dynamic fracture mechanics
– May 15, 2014
• Dynamic instabilities during fast fracture
– May 21, 2014

2 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 How strong is a solid? – 1. An atomic point of view

 2 (a  a0 ) 
Repulsion

F  Fmax sin  
  
Energy, E

a0
 2 a0  (a  a0 )
Interatomic    max sin   ,  
distance, a
   a0
Attraction

d 2 a0  2 a0 
E   max cos  
d      0
Tension

 0

2 a0 E  E
  max  max  ~
Force, F

  2a0 
a0 Interatomic
Compression

distance, a Modulus Strength

Material Strength/Modulus
GPa GPa
Steels 200 0.10 – 2 0.0005 – 0.01
Glass 70 0.17 0.0025
Carbon fibers 400 4 0.01
Glass fibers 70 11 0.16

Macroscopic strength is significantly smaller than the theoretical strength

3 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 How strong is a solid? – 2. The role of defects
1

 a a 2
 22 (a, 0)   1  2    22 (a, 0)  2   where   b 2 / a
 b 
1
 f a   max   constant Griffith’s experiments
2
 ~ 1E  10 m  1. Used experiments on glass
   f ~ 0.005 max tubes and glass bulbs loaded
a ~ 1E  6 m  under internal pressure to
show that 𝜎 𝑎 was constant
2. Manufactured fresh glass
x2
fibers with diameters in the
range of 1 mm to 3 microns
to show that small fibers had
2b strength of about 11 GPa
x1
2a

4 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 The continuum view of fracture

Lp

• Process of fracture can be cleavage, intergranular/ transgranular

fracture (polycrystalline materials), cavitation (ductile metals),
disentanglement (polymers), microcracking (glasses,…), fiber
breakage,…
• Details of processes within Lp are not important; only the total energy
needed for the fracture process is assumed to play a role in the
development of the fracture
• Lp is small – “small-scale process zone” – what does this mean?

5 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 The energy balance – a continuum view
Total energy of the system: E    U s
where U s surface energy (or fracture energy)
  W R  U R potential energy of the body
W R work done by the external forces on the body
UR strain energy stored in the body
d
G (a)   Energy Release Rate
da
define
dU s
R fracture resistance x2

da
At equilibrium, E   a   0 
x1

G  ac   R Fracture criterion 2a

Stable if E   ac   0
6 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Remarks 1: G  ac   R Fracture criterion
• ac is the equilibrium crack length (reversible)
• Fracture resistance R(a) includes the effect of all
dissipative fracture processes and is typically
calibrated from experiments.
• For Linearly Elastic Fracture Mechanics (LEFM),
the region outside Lp must exhibit linear elastic
behavior, but this is not a general requirement.
• Other than Lp being “small”, there is no length scale
here! The theory works at length scales from the
atomic to the tectonic.
• Need methods to calculate G(a) for specific crack
problems

7 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Remarks 2: G  ac   R Fracture criterion
• R varies over several orders of magnitude
– True surface energy is ~ O(1) J/m2
– Glasses and ceramics~ 10 J/m2
– Polymers ~1 kJ/m2
– Metals ~ 100 kJ/m2
• Differences arise due to different mechanisms of
deformation and failure
• Must be determined through calibration
experiments, such as the pioneering work of
Obreimoff (1930)

8 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Calculation of G(a)
CM compliance of the loading system P, T
CM  0  fixed displacement
CM    fixed load CM
C (a)   / P compliance of the specimen P, 
T total displacement (fixed)
 T    CM P   C ( a )  CM  P
a

1 1
Strain Energy: U R  CM P  C (a) P 2
2

2 2
d 1 2 Stable if E   ac   0 
  G (a )  P C (a )
da T 2
    C (a)  C (a)  CM 
2
C ( a )
1 2
P C (a )  R Stability depends on CM !
2
Can cause stick-slip and other unstable
equilibrium crack length crack growth effects
9 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Example 1: G(a) for a double cantilever beam
P , T

CM P, 

P,  d

2d
a

a
4a 3 12a 2
C (a)  ; C (a) 
3 2
8a 24a
C (a)  3
; C (a)  Ed 3
Ed 3
Ed Ed 3
1
ER
G  ac   R a P  2 2
 3Ed 3  2  4
12d 3 G  ac   R  a   
 8R 
unstable  C (a)  C (a )  C (a )  CM   stable
2

10 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Example 2: G(a) for an infinite strip specimen

E
G a 
1  2  h a
2h

 stable 

11 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Fracture mechanics – the global point of view
• The global point of view works quite well for a
number of problems.
• It circumvents detailed calculation of stress/strain
states in the vicinity of the crack.
• Has been applied successfully in a number of
structural applications
• Difficulty in calculating the compliance, C(a)
• Difficulty in calibrating the fracture energy, R
• Difficulty in selecting/identifying fracture path
• Modern numerical simulations incorporate the
energy approach through the phase-field
methodology.
12 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Fracture mechanics – why a local point of view?
• Provides a systematic way of calculating G(a)
• Provides a method for analyzing different loading
symmetries
• Local approach based on stress and strain fields
permits decoupling of path selection from failure
characterization

13 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Loading symmetries

Mode I or Mode II or Mode III or

Opening mode In-plane shear Anti-plane shear

14 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Linear elasticity
y ( x)  x  u ( x) Plane strain
1 u  u ( x1 , x2 ); u3  x3
ε(x)  u  (u)T 
2 
  ( x1 , x2 )   u ,   u , 
1
σ (x)   kk 1  2  ε  2
 σ  f  0  33  const;  3  0
- boundary conditions   ( x1 , x2 )      2 
u(x)  u *(x) on 1 R 
 33 ( x1 , x2 )   
s(x)  σ (x)n  s *(x) on  2 R 
 3 ( x1 , x2 )  0
Anti-plane shear
u  0; u3  u3 ( x1 , x2 ) If  11  ,22 ;  22  ,11 ;  12  ,12
then  4  0
 2u3  0

15 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 The J-integral

 u  x2
J    U R n1    n ds
 x1  n1


1. The integral is zero if contour is closed x1

inside the body without enclosing
singularities R 
2. If the contour goes from below to
above the crack surface as indicated,
the integral is independent of the path
3. This integral can be interpreted in
terms of the energy release rate:
J  G(a)  d  / da
4. Path independence implies that σ : ε ~ r 1
and therefore,

σ ~ r 1/ 2σ   ; ε ~ r 1/ 2ε  
16 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Anti-plane shear

K III
 3  h 3 ( )
2 r

K III  lim 2 r 32  r , 0  mode III stress intensity factor

r 0

1. Anti-plane shear can exist only in

specimens without bounding planes –
in axisymmetric geometries or Mode III or
infinitely thick plates Anti-plane shear
2. Free surfaces in finite thickness plates
introduce coupling to mode II
3. Connection to J and G obtained by 1 2
using the path independent integral: J  G (a )  K III
4. Failure criterion for mode III is still
2
being debated (more on this later!)

17 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 In-plane loading symmetries

Mode II or
Mode I or
In-plane shear
Opening mode

mode I stress intensity factor mode II stress intensity factor

K I  lim 2 r 22  r , 0  K II  lim 2 r 12  r , 0 
r 0 r 0

KI
   r ,   FI    T 1 1
2 r T represents nonsingular stress
K II
 FII    and plays a role in crack path stability
2 r

18 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Calculation of the stress intensity factors
• Elastic boundary value problem to be solved
– Numerous examples exist in handbooks
– Robust numerical methods based on FEM, BEM,
available
– Considered a solved problem: Given a geometry, loading,
etc, there is no difficulty in determining KI, KII, and KIII.
K I  K I  load, crack length, geometry 
P, 

• Example: Single-edge-notched specimen

P  a
KI  a f  
tW W  a

P a w
 g  displacement
tWE W 
19 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Fracture criterion for in-plane loading

1  2
1. Connection to J and G obtained by
using the path independent integral:
J  G (a) 
E
 K I2  K II2 

2. For pure mode I loading, the crack

grows along the line of symmetry and ER
the energy based fracture criterion
K I  K IC 
can be restated in terms of the stress
1  2
intensity factor

3. Residual strength (load carrying

capacity) can be determined for
structural applications

4. Stability of structures can be

evaluated

20 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Snap-back instability
P  a
KI  a f   P, 
tW W 
P a
 g   displacement
tWE  W 

a
a1
w

a2  a1

Crack is unstable in load control and displacement control

21 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Fracture criterion for mixed mode I + II crack

1. For combined modes I and II, we

need other criterion (criteria?) that
dictates the crack path selection
a. Maximize energy release rate
b. Maximum “hoop” stress
c. Principle of local symmetry:

K I  K IC , K II  0 Crack tilting

2. Maximum hoop stress criterion is

simplest to use
3. Experimental scatter is large and
unable to discriminate between the
different criteria

Crack kinking
Principle of Local Symmetry: Goldstein and Salganik, Int J Fract, 1974
22 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS
 Crack path evolution under Mode I + II

Yang and Ravi-Chandar, J Mech Phys Solids, 2001

Photograph Courtesy of Dov Bahat

Ben Gurion University
Tectonofractography, Springer

Principle of Local Symmetry works very well for this problem

23 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Possible fracture criteria for mixed mode I + III
Criterion I: Goldstein and Salganik,
Int J Fract, 1974

K II  0
f  K I , K III   0

Criterion II: Lin, Mear and Ravi-

Chandar, Int J Fract, 2010

K I  K IC , K II  0, K III  0

1  K III
    tan 2  
2  KI
Hull, Int J Fract, 1995
Cooke and Pollard, J Geophy Res, 1996

24 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Mixed mode I + III crack problem

Below a threshold of KIII/KI, the crack front twists

cr= 3.3°
Above the threshold, crack front fragments
Sommer, Eng Frac Mech, 1970

25 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Mixed mode I + III crack problem

Knauss, Int J Fract, 1971

26 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS

 Summary and plan
• Energy based method can provide a simple way of
analyzing fracture problem (with some residual
difficulty regarding the path selection)
• Stress-intensity factor based method provides an
effective way of designing fracture critical
structures – residual strength diagram

27 CENTER FOR MECHANICS OF SOLIDS, STRUCTURES AND MATERIALS