You are on page 1of 34

CEE 770 Meeting 6 Objectives of This Meeting

Learn theories for predicting mixed mode interaction, crack trajectory in 2D, and stability of such trajectory: 1st Order, LEFM theories, isotropic material Crack kinking vs crack turning: trajectory stability 2nd Order, LEFM theory, isotropic and orthotropic materials

102

Observation and Your Thoughts?


Cracks do not usually propagate as straight lines, or flat surfaces, or perfect ellipses.

What controls the shape (sometimes called trajectory when 2D idealization is reasonable) of a propagating crack?

Why do some cracks in a symmetric structure with symmetric BCs not propagate symmetrically?

103

Examples of These Observations


Very non-simple crack shapes!

Symmetric structure, BCs: Unsymmetric crack growth??


104

One of My Favorite Quotes.


... NOTHING AT ALL HAPPENS IN THE UNIVERSE IN WHICH THERE DOES NOT SHINE OUT SOME PRINCIPLE OF MAXIMUM OR MINIMUM, WHEREFORE THERE IS ABSOLUTELY NO DOUBT BUT THAT ALL HAPPENINGS IN THE UNIVERSE MAY BE DETERMINED FROM FINAL EFFECTS BY A METHOD OF MAXIMA OR MINIMA QUITE AS SUCCESSFULLY AS FROM ACTUAL CAUSES THEMSELVES. L. EULER, 1744

105

Recall Continuum Fracture Modes


y,v x,u y,v x,u y,v x,u

z,w

z,w

z,w

Mode I

Mode II

Mode III

Basic modes of crack loading. Positive sense shown for each: Mode I = crack opening Mode II = in-plane sliding Mode III = anti-plane tearing
106

And, the four shell fracture modes


h 2

z y r

membrane
KI K II

bending

K1 k1

Reissner theory Kirchhoff theory

K2 , K3 Reissner theory k2

Kirchhoff theory

107

1st Order, LEFM Crack Kinking Theories


1st Order LEFM theories are based on only the singular terms of the local asymptotic LEFM crack front fields. Many such theories have been proposed and tested, and most of these are variants of these 3:
Maximum Hoop Stress Theory

max

Erdogan and Sih (1963) Hussain et al. (1974) Sih (1974)

Maximum Energy Release Rate Theory G()max Minimum Strain Energy Density Theory

S()min

We will study only the max theory, here, but will return to the concept of maximum energy release rate theory later. Why, and why?
108

1st Order, LEFM, Isotropic Crack Kinking Theories: max Theory


Recall equation 16, p.36:
=
1 3 T cos K I cos2 K II sin + (1 cos 2 ) 2 2 2 2r 2

(16)

This theory asserts that, for an isotropic material, a crack will kink into the direction normal to the maximum circumferential (hoop) stress. So maximize (16) wrt , ignoring T-stress, set = c, and rearrange,
y

sin c K II = K I (3 cos c 1)
max max

Then solve for c


c

1 1 + 8(K K )2 II I c = 2tan 4( K K ) II I
1

(65)

109

Mixed-Mode Interaction According to max Theory


Rewrite (16), again neglecting T-stress, and recognize that the new left-hand side represents a measure of fracture toughness that, in the limit of Mode I only, must be KIc

3 1 cos K I cos2 K II sin 2 2 2 2r


(66)

3 rc = cos K I cos2 K II sin = K Ic 2 2 2

Equations 65 and 66 comprise a parametric set in c, KI, and KII .These can be solved to produce an interaction diagram that is analogous to a multi-axial yield interaction diagram, or a biaxial bending yield-crushing diagram.

110

Mixed-Mode Interaction According to max and Other Theories


1.2 1
Note that each theory has its own interaction surface, and its own Mode II toughness prediction

S min ( = 0.25) max G max K eff

0.8 K / K c 0.6 0.4 0.2 0

0 0.2 0.4 0.6 0.8 1


111

= tan-1(KII/KI)

/ c

Another Way To View Effective, Mixed-Mode Toughness


1.2 1

S min ( = 0.25)

K( ) / K Ic

max

0.8 0.6 0.4 0.2 0

max

max

G max

60 ,

90

(b)

30 60 Load Angle, = tan-1(KII/KI)

90

112

Comparison of 1st Order, Linear Elastic, Isotropic Crack Kinking Theories: Kink Angle

Mode II Mode mixity parameter: M e =


1 K tan I KII 2

Mode I

113

Demonstration of the Maximum Hoop Stress Crack Turning Criterion


SEN(B) polymethylmethacrylate (PMMA) beams. Initial crack location and length were varied among the specimens.
P Note: all dimensions in inches 2.0 0.5 dia. typ. thickness: 0.5 8.0 2.0

2.75

4.0

a 9.0 10.0 9.0 10.0

114

Comparisons between observations and predictions for two different initial crack configurations

This is VG predicting! How big is process zone in PMMA?


analysis experiment Analysis crackincrement lengths: a = 0.3 inch a = 0.2 inch a = 0.05 inch pre-cut slot 1.0 inch 6.0 inches (from centerline) 2.5 inches (from bottom of plate) pre-cut slot 6.0 inches (from centerline)

115

Crack Kinking versus Crack Turning


Crack path problems encountered in most real structural applications are not really crack kinking problems. In an average macroscopic sense, cracks typically propagate in a rather smoothly turning fashion as the crack negotiates its way among the structural features of the part. Since the first-order isotropic theories predict crack kinking for non-zero KII , the only way for a crack to propagate smoothly is for the crack to follow a path along which KII=0. Since all the first-order isotropic theories agree exactly for this condition, the crack path is apparently independent of any first-order theory.

116

Then Why Do Some Mode I Cracks Turn?

There appears to be some trajectory instability phenomenon at work.


117

Then Why Do Some Mode I Cracks Turn? Crack Turning Theory


Consider a Mode I crack subjected to a small trajectory perturbation at x= 0, i.e. the crack propagates very slightly out of its self-similar direction and feels some small, corresponding KII. Also, lets include the first 2nd order field term, the T-stress. Cotterell and Rice (1980) then asked: What happens to continuing trajectory if we enforce the condition that subsequent KII=0?

0
2 K II KI

Strength of the perturbation: 0 =

Strength of the T-stress: = 2 2

T KI
118

Cotterell and Rice Crack Turning Theory


Cotterell and Rice found that subsequent trajectory is
(x) = o 2 x 2 ( ) exp x erfc x 1 2

=2 2

T KI

Trajectory unstable for Positive T-stress !

0 =

2 K II KI

Normalized Plot of the Perturbed Crack Path of Cotterell and Rice (1980).
119

Many Materials Exhibit More Complicated Behavior Such as Toughness Orthotropy and Crack Path Sensitivity to Load Level

Objective: develop a theory for crack turning in real materials based on LEFM concepts
120

Material Orientation Definitions for Fatigue and Fracture

121

2nd Order Theories: Role of T-Stress


Cotterell and Rice (1980) crack perturbation theory highlighted the importance of the T-stress in trajectory predictions.

Their work inspired the creation of 2nd order theories for prediction of crack shape.

We will investigate one of these 2nd order theories, and extend our thinking about crack shapes to the more general case of materials with anisotropic toughness.

122

Recall T Is Second Term of the Crack-tip Stress Expansion


3 K I (1 + sin 2 ( 2) ) + 2 K II sin( ) 2 K II tan( 2) rr 1 + cos(2 ) 1 T 3 cos K I cos2 ( 2) 2 K II sin( ) = + 1 cos(2 ) 2 2r 2 sin(2 ) sin( ) ( 3 cos( ) 1 ) K K I II r

Include the T-term and the maximum hoop stress expression then becomes:

K II 2 sin( 2) 8 T = cos 2rc cos KI 3cos 1 2 3 KI


rc is the distance from the crack tip at which the stresses are computed. rc scales with the plastic zone size. rc for plastic tearing is theorized to be a material constant. T y

(67)

Kosai, Kobayashi, and Ramulu, Tear straps in aircraft fuselage, Durability of metal aircraft structures: Proc. of Int. Workshop Structural Integrity of Aging Airplanes, Atlanta, GA, 443-457, 1992

123

2nd Order Linear Elastic, Crack Turning Theory, Isotropic Case

Normalized Crack Turning Plot for Isotropic Material Based on the Formulation of Kosai et al. (1992).

124

Crack turning interaction diagram


80 60 40 20
90o 67.5o 45o 22.5o 5o 5o 22.5o 67.5o 45o

max criterion

tan 1 (K II K I )

0
-20 -40 -60 -80 -4 -3 -2 -1

1o 1o

tan 1 (K II K I ) = 0o

tan 1 (K II K I ) = 90o

8T 3K I

2rc

Pettit, Wang, and Toh, Integral airframe structures (IAS) - validated feasibility study of integrally stiffened metallic fuselage panels for reducing manufacturing cost, Boeing Report CRAD-9306-TR-4542, NASA contract NAS1-20014, Task 34, November, 1998.

125

Conceptual Model of a Crack Propagation Criterion for a Toughness Orthotropic Structure


c
Predicted direction of crack propagation

rc
() evaluated at rc

Material toughness function

( K I , K II , E 1 , E 2 , T , rc , ) Maximum = K c ( ) Kc
Vol. 35 (1990) pp. 159-170

critical

(68)

Boone, Wawrzynek, and Ingraffea, Analysis of fracture propagation in orthotropic materials, Engng Fracture Mech,

126

A Simple Representation for Toughness Anisotropy: The Toughness Ellipse

Kp is the stress intensity at which the crack propagates, in the relevant regime of crack growth. Thus, for fatigue crack growth, Kp is the stress intensity at which the crack propagates at a given rate; for stable tearing, Kp represents the fracture toughness.

2 cos2 sin K p ( ) n + n =1 K p (90) K p (0)


n

K p (90) Km K p (0)
127

Typical 2nd Order Interaction Diagrams for Orthotropic Toughness


1.2

100 75 50 25

crack oriented at = 0

Kc L T = 1.2 Kc T L
1.0

0
-25 -50 -75 -100 -4 -3 -2 -1 0 1 8T 2r c 3K I 2 3 4

crack

ellipse for anisotropic fracture resistance


100 75 50 25

crack oriented at = 45

100 75 50 25

crack oriented at = 90

0
-25 -50 -75 -100 -4 -3 -2 -1 0 1 8T 2r c 3K I 2 3 4

0
-25 -50 -75 -100 -4 -3 -2 -1 0 1 8T 2r c 3K I 2 3 4

128

Normalized Crack Turning Plots for an Elastically Isotropic Material with Fracture Orthotropy Km =1.6, n = -1, Various Crack Orientations.

(a) Crack Oriented at =0, (b) Crack Oriented at =45, (c) Crack Oriented at =90.

129

rolling direction (L) (T)

K I = 70 ksi in. K II = 0 T = 50 ksi rc = 0.05 in. K c (T ) K c ( L) = 1.1

No T-Stress

T-Stress

T-Stress and Orthotropy

Propagation direction

c = 0o

c = 23.5o

c = 45.6o

130

Observed and predicted crack paths for 7050-T7451 DCB specimens, Static Loading
2.5 2.0 1.5 1.0 0.5 0.0 -0.5 0 1 2 3 4 5 6 Horizontal Crack Growth (in) 7 8 9 10 rc=0 , Km = 1.3 rc=.05 inches, Km = 1.3 rc=.1 inches, Km = 1.3 rc-LT-15-5

2.5 Vertical Crack Growth (in) 2.0 1.5 1.0 0.5 0.0 -0.5 0 1 2 3 4 5 6 Horizontal Crack Growth (in) 7 8 9 10
rc-TL-15-5 rc=.05 inches, Km = 1.3 rc=0, Km = 1.3

131

Observed and predicted crack paths for 7050-T7451 DCB specimens, Fatigue Loading
2.5

Vertical Crack Growth (in)

2.0 1.5 1.0 FRANC2D, Km=1.1, rc=0 0.5 0.0 rc-LT-15-2 rc-TL-15-2

-0.5 0 1 2 3 4 5 6 7 8 9 10

Horizontal Crack Growth (in)

132

What About These Data? What is Going on Here?


100 80 60 40 c 20 0 -20 -40 0 10 20 30 40 50 60 70 Mode mixity, - Mode II Dominated 80 90 Mode I Dominated LEFM Max stress Pure mode virtual kink, Eq. (5.30) Curve fit to 2024-T3 Test Data [40] SSY CTOD Analyses [40] 2024-T3 Arcan Test Data [40]

Look only at the test data and the LEFM Max stress, max , information.

There is an obvious, abrupt change in trajectory behavior. Why?

133

Predicted Effect of T-Stress on Kink Angle for Mode II Crack According to Maximum Shear Stress Theory, Isotropic Case.

How would you formulate such a theory?

134

Predicted Effect of T-Stress on Kink Angle for Mode II Crack According to Maximum Shear Stress Theory, KII m=1.6, n=-1

(a) Crack Oriented at = 0

(b) Crack Oriented at = 90

135