Force Vectors (Tensors)

• A force, F , is a vector (also called a "1st-order tensor")

• The description of any vector (or any tensor) depends

on the coordinate system used to describe the vector:
100 lb 100 lb 100 lb

+y
+x"
+x'
+y'
+y" 60 deg
45 degs

+x
F = 100 lb F = 70.7 lb (x'-direction) F = 86.6 lb (x"-direction)
(y-direction) + +
70.7 lb (y'-direction) 50 lb (y"-direction)
Normal and Shear Forces

• A "normal" force acts perpendicular to a surface

• A "shear" force acts tangent to a surface

P = Normal Force
V = Shear Force
Forces Inclined to a Plane

• A force inclined to a plane can always be described

as a combination of normal and shear forces

Inclined Force
P = Normal Force
V = Shear Force
Moments

• A moment (also called a "torque" or a "couple") is a

force which tend to cause rotation of a rigid body
• A moment is also vectoral quantity...

M
 Static Equilibrium
• A rigid solid body is in "static equilibrium" if it is:
– at rest, or
– moves with a constant velocity
• Static equilibrium exists if:  F = 0 and  M = 0
50 lbf 50 lbf

(40 lbf) (40 lbf)

(30 lbf) (30 lbf)

60 lbf

(30 lbf) (BALL ACCELERATES) (30 lbf) (NO ACCELERATION)

(40 lbf) (40 lbf)

50 lbf 50 lbf
Free Body Diagrams and "Internal Forces"

• An imaginary "cut" is made at plane of interest

• Apply  F = 0 and  M = 0 to either half to
determine internal forces, R
F F

R (= F)

"cut"
(or)

R (= F)

F F
Free Body Diagrams and "Internal Forces"
• The imaginary cut can be made along an arbitrary
plane
• Internal force R can be decomposed to determine the
normal and shear forces acting on the arbitrary plane
F

R (= F) P

"cut" V

F F F
Stress
Fundamental Definitions
• Two "types" of stress:
– normal stress =  = P/A
– shear stress =  = V/A
– where P and V must be uniformly distributed over A

P = Normal Force  = P/A

V = Shear Force  = V/A

A = Cross-Sectional Area
Distribution of Internal Forces

• Forces are distributed over the internal plane...

they may or may not be uniformly distributed

F M

"cut" "cut"

F F M M
 Infinitesimal Elements
• A free-body diagram of an "infinitesimal element" is
used to define "stress at a point"
• Forces can be considered "uniform" over the
infinitesimally small elemental surfaces
+x

+z
+y

dx

dy
dz
Labeling Stress Components

• Two subscripts are used to identify a stress

component, e.g., “xx” or “xy” (note: for convenience
we sometimes write x = xx or xy = xy)


yx

+y  xy
 
xx +x xx
 xy

yx
Admissable Pure Shear Stress States

yx

+y
 +y
 +y

xy xy xy

+x
 +x
 +x
xy xy

yx

If : τyx = τxy
F  0 F = 0
then :
M  0 M  0
(inadmissable) (inadmissable) F = 0
M = 0
(admissable)
Stress Sign Conventions

• The algebraic sign of an element face is “positive”

of the outward-pointing unit normal to the face
“points” in a positive coordinate direction

• A stress component is positive if:

– stress component acts on a positive face and “points” in
a positive coordinate direction, or
– stress component acts on a negative face and “points”
in a negative coordinate direction
Stress Sign Conventions

yy yy

+y  xy +y
 xy
xx 
+x xx
+x

all stresses positive xx and yy positive,

xy negative
 The Stress Tensor
• Stress is a "2nd-order tensor", and in the most general
case six components of stress exist "at a point"
_
F1 _
F2 xx +x
xy
xz
yx
+y
yz
_ zy
F3
zz yy
_
F5
_ zx
F4 +z
 Plane Stress
• If all non-zero stress components exist in a single
plane (i.e., if zz = xz = yz = 0), the state of stress
is called "plane stress"
+y +y

yy yy
+x xy
+z xy
xx
xx
xx xx
+x

yy
yy
 Uniaxial Stress
• If only one normal stress exists (if xx = zz = xy
= xz = yz = 0), the state of stress is called a
"uniaxial stress"
+y
+y

yy yy
+x
+z

+x

yy
yy
Free Body Diagram Defines the
Coordinate System

F
+y
+y

F yy
"cut"
+x +x

F F F
Free Body Diagram Defines the
Coordinate System
F
y' y'
y'y'
P V x'y'
"cut" x' x'

F F F
Free Body Diagram Defines the
Coordinate System

y" y"
"cut" y"y"
(a Py"y"
plane) x"y"
x" x"
Vy"x"
Vy"z" y"z"
z" z"

F F F
Stress Transformations

• Given stress components in the x-y coordinate

system (xx, yy, xy), what are the corresponding
stress components in the x'-y' coordinate system?
yy
+y
y'y'
x'y'
xy +y' x'x'
xx
? +x'


+x
 Stress Transformations

• Stress components in the x'-y' coordinate system may

be related to stresses in the x-y coordinate system using
a free body diagram and enforcing  F = 0

+y yy Fx' = 0
+y' +x'

xy x'y' 

xx xx x'x'

+x xy
yy
"cut"
Stress Transformation Equations

• By enforcing Fx' = 0, Fy' = 0, it can be shown:

 x'x' =  xx cos 2  +  yy sin 2  + 2 xy cos sin 

 y' y' =  xx sin2  +  yy cos2  − 2 xy cos sin

 x' y ' = ( yy −  xx )cos sin  +  xy (cos 2  − sin 2  )

Important note: angle  is positive from the +x-

direction towards the +y'-direction (for the
elements drawn in this presentation, a positive
angle  is counter-clockwise)
Stress Transformation Equations

• Using matrix notation, these can also be written:

 x'x'   cos 2  sin 2  2 cos sin    xx 

    
 y' y'  =  sin 
2
cos 2  − 2 cos sin    yy 
  − cos sin  cos sin (cos 2  − sin 2  )   xy 
 x' y'  
Stress Transformation Equations
• These transformation equations can also be
visualized using Mohr’s circle of stress:
 yy
+y
xy

xx 
(yyxy) (x'x'−x'y')
+x

2 2
(Equivalent)

 y'y'
x'y' ( y'y'  x'y' ) (xx−xy)
+y
x'x'
'
+x'

 Principal Stresses

• In the principal stress coordinate system the shear

stress is zero, and the normal stresses are max/min...
 xx +  yy   xx −  yy 
2

11 ,  22 =    +  xy2
2  2 
1  2 xy 
−1 
 p = tan  

2   xx −  yy 

*** OR, EQUIVALENTLY ***

  11 −  xx 
 p = tan 
−1 
  
 xy 
 Maximum Shear Stress

• In the maximum shear stress coordinate system the

in-plane shear stress is at a maximum and in-plane
normal stresses are equal
 +
11 +  22
 x =  y = xx yy
=
2 2
  xx −  yy 
2
m =    +  xy
2
 2 
1 −1   yy −  xx 
t = tan
2  2 xy 
 
“Transformation” of Stress
yy y'y'
+y x'y'
xy +y'
 x'x'
(Equivalent)
 xx
+x'
'
+x

+y"
 y"y"

x"y"
11
x"x" "
p
+x"
 Sample Problem

• Given the following stress components (where the x-axis

is horizontal and positive to the right, and the y-axis is
vertical and positive upwards):
xx = 3500 psi yy = -30000 psi xy = 20000 psi
(a) Sketch the stress element in the x-y coordinate system
(b)Sketch the stress element in the x'-y' coordinate
system, oriented 50°CCW
(c)Sketch the stress element in the principal stress
coordinate system
(d)Sketch the stress element in the maximum shear stress
coordinate system
 Sample Problem (answers)
+y 30000 psi

20000 psi

Part (a) Part (b)

3500 psi

+x

Part (c) Part (d)

Sample Problem
Use of Stress Transformation Equations (example)

 x'x'   cos 2  sin 2  2 cos sin  3.5ksi 

    
 y' y'  =  sin 
2 cos 2  − 2 cos  sin  − 30ksi 
  − cos  sin  cos  sin  (cos 2  − sin 2  ) 
 x' y'    20ksi 

 x'x'   cos2 50 sin2 50 2 cos 50 sin 50  3.5ksi 

    
 =
 y' y'   sin 2 50 cos 2 50 − 2 cos 50 sin 50  − 30ksi 
  − cos 50 sin 50 cos 50 sin 50 (cos 2 50 − sin 2 50) 20ksi 
 x' y'    
 x'x'   3.538ksi 
   
 y' y'  = − 30.038ksi 
  −19.968ksi
 x' y'   
Sample Problem
Use of Mohr’s Circle
 "Stress": Summary of Key Points

• Normal and shear stresses are both defined as a (force/area)

• Six components of stress must be known to specify the "state
of stress" at a point
• Stress is a tensorial quantity; values of individual stress
components depend on the coordinate system used
• Stress is defined strictly on the basis of static equilibrium;
definition is independent of:
– material properties
– strain
– temperature
Strain
Fundamental Definitions
• "Strain" is a measure of the deformation of a solid
body
• There are two "types" of strain; normal strain () and
shear strain ()

 = (change in length) units = in/in, m/m, etc

(original length)

 = (change in angle) units = radians

 The Strain Tensor

• Strain is a 2nd-order tensor, and in the most general

case, six components of strain exist "at a point":
xx, yy, zz, xy, xz, yz

• Since strain is a tensorial quantity, the values of the

individual strain components which define the "state
of strain" depend on the coordinate system used…

• This review will primarily involve strains which

exist within a single plane
Strain Within a Plane

• We often encounter two distinct conditions that

result in problems involving “strains within a plane”:
– Plane Stress: All non-zero stress components lie within a
single plane (e.g., xx, yy, xy  0 , zz = xz = yz = 0). If
the material is isotropic, the plane stress condition induces
four non-zero strain components: xx, yy, zz, and xy

– Plane Strain: All non-zero strain components lie within a

single plane (e.g., xx, yy, xy  0 , zz = xz = yz = 0). By
definition, the plane strain condition involves three non-
zero strain components: xx, yy, and xy
Strain Within a Plane
 Strain Sign Convention

• A positive (tensile) normal strain is associated with an

increase in length
• A shear strain is positive if the angle between two
positive faces (or two negative faces) decreases

+y +y

+x +x
All Strains Positive xx Positive
 yy and  Negative
xy
Visualization of Strain
(Assuming link made of an isotropic material)
F

+y (Loading)
yy > 0
+x xx < 0
xy = 0

F
Visualization of Strain
(Assuming link is made of an isotropic material)
F

+y' +x' (Loading)

x'x' = y'y' > 0

x'y' > 0
45 deg

F
Strain Transformations

• Given strain components in the x-y coordinate

system, what are the corresponding strain
components in the x'-y' coordinate system?

+y +y'
+x'

?
+

+x
Given: xx, yy, and xy Find: x'x' , y'y', and x'y'
 Strain Transformation Equations

• Based strictly on geometry, it can be shown:

 
 =  cos  +  sin  +  xy 2 cos sin 
2 2
x'x' xx yy
 2 
 
 =  sin  +  cos  −  xy 2 cos sin 
2 2
y' y ' xx yy
 2 
 x' y'   xy  2 2
= ( yy −  xx ) cos sin  +  (cos  − sin  )
2  2 
Important note: angle  is positive from the +x-direction
towards the +y-direction (counterclockwise as drawn in
this presentation)
Well, I'll Be Darned!!!

• The stress transformation equations are based

strictly on the equations of static equilibrium

• The strain transformation equations are based

strictly on geometry

• Nevertheless, the stress and strain transformation

equations are nearly identical!! (…because both
stress and strain are 2nd-order tensors…)
Strain Transformation Equations
• Using matrix notation, these can also be written

  x'x'   cos 2  sin 2  2 cos sin     xx 

    
 y' y'  =  sin 
 2
cos 2  − 2 cos sin     yy 
 / 2 − cos sin  cos sin (cos 2  − sin 2  )  xy / 2
 x' y'  
Strain Transformation Equations

• These transformation equations can also be

visualized using Mohr’s circle of strain:
+y


(yyxy/2)
(x'x'−x'y'/2)
+x

2
(Equivalent) 2

(xx− xy/2)
(y'y'  x'y' /2)
+y'
+x'

 Principal Strains

• In the principal strain coordinate system the shear

strain is zero and the normal strains are max/min...
 xx +  yy   −  yy    xy 
2 2

11 , 22 =   xx  + 
+2 2  2   2 
(11, 22) 1   xy 
 p = tan−1 
2  − 
 xx yy 

*** OR, EQUIVALENTLY ***

p
+1

 2(11 −  xx ) 
 p = tan −1  
  xy 
 
“Transformation” of Strain

(xx, yy, xy)

+y (x'x', y'y', x'y' )
+y'

(Equivalent) +x'
'
+x

+y"
+2 (x"x", y"y", x"y" )
(11, 22)

"
p
+1 +x"
 "Strain": Summary of Key Points
•  = ( length)/(original length)  = ( angle)
• Six components of strain specify the "state of strain"
• Strain is a tensorial quantity; numerical values of individual strain
components depend on the coordinate system used
• Strain is defined strictly on the basis of a change in shape;
definition is independent of:
– material properties
– stress
– temperature
• "Suprisingly," the stress and strain transformation equations are
nearly identical
 Sample Problem
• Given the following strain components (where the x-
axis is horizontal and positive to the right, and the y-axis
is vertical and positive upwards):
xx = 2000 in/in yy = -1350 in/in xy = 2200 rad

• (a) Sketch (not to scale) the strain element in the x-y

coordinate system
(b)Sketch (not to scale) the strain element in the x’-y’
coordinate system, oriented 50°CW
(c)Sketch (not to scale) the strain element in the
principal strain coordinate system
 Sample Problem (answers)
+y +y'
+2

16.6o

o
-50
+1
+x

+x'

Part (a): Part (b): Part (c):

xx = 2000 in/in x’x’ = -1049 in/in 11 = 2329 in/in
yy = -1350 in/in y’y’ = 1699 in/in 22 = -1679 in/in
xy = 2200 rad x’y’ = 2917 rad p = 16.6°
Hooke’s Law

• The structural engineer is typically interested in

the measuring the state of stress induced in a
structure during service
• The state of stress cannot be measured directly….
• The state of strain can be measured directly….
• Hence, we must develop a relationship between
the stress tensor and the strain tensor…this
relationship is called a “constitutive model”, and
the most common is “Hooke’s Law”
 Hooke’s Law (Cont’d)

• The form of Hooke’s depends on whether the

material is isotropic or anisotropic:
– Isotropic materials: same properties in all directions
– Anisotropic materials: properties vary with direction
• More than one “type” of anisotropic behavior. Three
will be mentioned in this review:
– Transversely isotropic
– Orthotropic
– Generally anisotropic
Isotropic vs Anisotropic Materials

Anisotropy occurs because of some type of order in

the microstructure
 Hooke’s Law

• Hooke’s Law will be reviewed/discussed in the

following order:
– Isotropic materials
– Anisotropic
• Transversely isotropic
• Orthotropic
• Generally anisotropic
 Isotropic Material Properties
The Uniaxial Tensile Test
• Specimen is subjected to axial tensile force, inducing a uniaxial
state of stress in the "gage" region
• Stress is increased until fracture occurs; corresponding axial and
transverse strains are measured throughout the test (all shear
strains = 0)
xx = l/l yy = zz = w/w = t/t xy = yz = xz = 0
The Tensile Stress-Strain Curve
A plot of axial stress vs axial strain


Tensile Strength

Fracture Stress

0.2% Offset
Yield Strength

0.2% Strain to Fracture 

 Load-Unload Cycles

  
Tensile Strength Tensile Strength Tensile Strength

0.2% Offset
0.2% Offset 0.2% Offset
Yield Strength
Yield Strength Yield Strength

  
Elastic Strain Elastic Strain
Inelastic
Inelastic Strain Elastic Strain
Strain
 Material Property:
Young’s Modulus
• At stress levels below the

yield stress the response is
called "linear elastic"
• The slope of the linear region 0.2% Offset
is called "Young's modulus" Yield Strength
or the "modulus of elasticity", In Linear Region:
E E  = E
• In the linear region and for a
uniaxial stress-state (only!!!):
 = E (or)  = /E 1


 Material Property:
Poisson's Ratio
• Poisson's ratio is based on the ratio of two normal
strains cause by a uniaxial stress:  = -(t/a)
• Poisson’s ratio is a measure of the coupling
between xx and yy, zz
• In this case:  = (-yy/xx) = (-zz/xx)
 Material Property:
Poisson's Ratio
• Poisson's ratio is based on the ratio of two normal
strains cause by a uniaxial stress:  = -(t/a)
• Poisson’s ratio is a measure of the coupling
between xx and yy, zz
• In this case:  = (-yy/xx) = (-zz/xx)
Material Properties
The Torsion Test

• Thin-walled cylindrical specimen subjected to a

torque, inducing a uniform shear stress xy in the
gage region of the specimen
• Shear stress (i.e., torque) increased until fracture
occurs; shear strain measured throughout test
+y

xy

+x

T
T
 The Shear Stress- Shear Strain Curve

• At linear levels, the slope of

the shear stress -shear strain 
curve is called the “shear
modulus”

• In the linear region (only!!)

xy = Gxy (or) xy = xy/G G


Number of Independent Material
Properties
• Three material properties have been defined; E, ,
and G
• For an isotropic material, only two of these three
properties are independent…it can be shown:

E
G=
2(1+  )
 Derivation of Hooke’s Law
For an isotropic material subjected to general 3D streses

• We assume the strain

+x tensor is linearly related to
F1
+z xx
the stress tensor….(when is
+y
M1
F2
xz this a bad assumption?)
zy
xy • Assuming the linear
zz
assumption is appropriate,
F3 yy the principle of
F4
M2 superposition can be used
to develop a Hooke’s law:
Hooke’s Law (cont’d)
Strains caused by xx only:

(What strains are induced

by xx only?)
Hooke’s Law (cont’d)
Strains caused by xx only:

xx = k11 xx = (1/E) xx

yy = k21 xx = (-/E) xx
zz = k31 xx = (-/E) xx
xy = k41 xx = 0
yz = k51 xx = 0
zx = k61 xx = 0
Hooke’s Law (cont’d)
Strains caused by yy only:

xx = k12 yy = (-/E) yy

yy = k22 yy = (1/E) yy
zz = k32 yy = (-/E) yy
xy = k42 yy = 0
yz = k52 yy = 0
zx = k62 yy = 0
Hooke’s Law (cont’d)
Strains caused by zz only:

xx = k13 zz = (-/E) zz

yy = k23 zz = (-/E) zz
zz = k33 zz = (1/E) zz
xy = k43 zz = 0
yz = k53 zz = 0
zx = k63 zz = 0
Hooke’s Law (cont’d)
Strains caused by xy only:

xx = k14 xy = 0

yy = k24 xy = 0
zz = k34 xy = 0
xy = k44 xy = (1/G)xy=[2(1+)/E] xy
yz = k54 xy = 0
zx = k64 xy = 0
Hooke’s Law (cont’d)
Strains caused by yz only:

xx = k15 yz = 0

yy = k25 yz = 0
zz = k35 yz = 0
xy = k45 yz = 0
yz = k55 yz = (1/G)yz =[2(1+)/E] yz
zx = k65 yz = 0
Hooke’s Law (cont’d)
Strains caused by zx only:

xx = k16 zx = 0

yy = k26 zx = 0
zz = k36 zx = 0
xy = k46 zx = 0
yz = k56 zx = 0
zx = k66 zx = (1/G)zx =[2(1+)/E] zx
Hooke’s Law (cont’d)
Strain xx caused by all stress components acting
simultaneously:

xx = ?
Hooke’s Law (cont’d)
Strain xx caused by all stress components acting
simultaneously:

xx = ?

(since strain-stress relation

assumed linear, we can
apply the principle of
superposition and simply
add up contribution of
each stress component):
Hooke’s Law (cont’d)
Strain xx caused by all stress components acting
simultaneously:

xx = k11 xx

+ k12 yy
+ k13 zz
+ k14 xy
+ k15 yz
+ k16 zx
Hooke’s Law (cont’d)
Strain xx caused by all stress components acting
simultaneously:

xx = (1/E) xx

+ (-/E)yy
+ (-/E) zz
+ (0) xy
+ (0) yz
+ (0) zx
Hooke’s Law (cont’d)
Rearranging:

T he picture can't be displaye d.

1

 xx =  xx − ( yy + zz )
E

Hooke's Law
Repeating process for all six strain components

2(1 +  ) xy
1

 xx = xx −  ( yy +  zz )
E
  xy =
E
2(1 +  ) xz
1

 yy =  yy −  ( xx +  zz )
E
  xz =
E
2(1 +  ) yz
1

 zz = zz −  ( xx +  yy )
E
  yz =
E
Hooke’s Law
Matrix Notation

xx  1 − − 0 0 0   xx 
  −  
  1 − 0 0 0
yy
   yy 
 zz  1 −  − 1 0 0 0   zz 
  =   
 yz  E  0 0 0 2(1+  ) 0 0
 yz 
 xz  0 0 0 0 2(1+  ) 0    xz 
  0   
 xy   0 0 0 0 2(1+  )  xy 
 Hooke’s Law
Inverting the six equations leads to a more convenient
form for experimental analysis…

E xy
 xx = E
(1 +  )(1 − 2 )

(1− ) xx +  ( yy +  zz )   xy =
2(1 +  )
E xz
 yy = E
(1 +  )(1 − 2 )
(1 −  ) yy +  ( xx +  zz )   xz =
2(1 +  )
E yz
 zz = E
(1 +  )(1 − 2 )

(1− ) zz +  ( xx +  yy )   yz =
2(1 +  )
 Hooke’s Law
Matrix Notation

(1−  )   0 0 0 
 xx    (1−  )  0 0 0   xx 
    
   (1−  )  
 yy  0 0 0 yy 
 zz  E  (1− 2 )   zz 
 =  0 0 0 0 0  
  yz  (1+  )(1− 2 )  2
  yz 
(1− 2 )
  xz   0 0 0 0 0  xz 
   2  
 xy   
(1− 2 )   xy 
0 0 0 0 0
 2 
Hooke's Law For Plane Stress
assume zz = xz = yz = 0
+y

yy
xy
xx
xx
+x

yy 2(1 +  ) xy
E

 xx = 1 xx −  ( yy +  zz )   xy =
E

 yy = 1  yy −  ( xx +  zz )   xz = 2(1 +  ) xz
E E
2(1 +  ) yz
 zz = 1 zz −  ( xx +  yy )  yz =
E E
Hooke's Law For Plane Stress
assume zz = xz = yz = 0
+y

yy
xy
xx
xx
+x
0
yy 2(1 +  ) xy
1

xx =  xx −  ( yy +  zz )
E
  xy =
E 0
0
 
 yy = 1  yy −  ( xx +  zz )  xz = 2(1 +  ) xz
E E 0
0
2(1 +  ) yz
 zz = 1 zz −  ( xx +  yy )  yz =
E E
Hooke's Law For Plane Stress
assume zz = xz = yz = 0

 xx = 1 ( xx −  yy )
+y E
yy
(
 yy = 1 yy − xx )
xy E
xx
xx  zz = −
E
(  xx +  yy )
+x
2(1 +  ) xy
yy  xy =
E
 xz =  yz = 0
Hooke's Law For Plane Stress
assume zz = xz = yz = 0

+y
 xx = E
 xx +  yy 
yy (1 −  2 )

xx
xy
 yy = E
  yy + xx 
xx (1 −  2 )
E
 xy
+x
 xy =
yy 2(1 +  )
 zz =  xz =  yz = 0
 Hooke's Law for Uniaxial Stress
If xx = zz = xy = xz = yz = 0
+y
yy

+x = 1
 0 0
2(1 +  ) xy
0

yy xx
E
xx −  ( yy +  zz )   xy =
E 0
0 0

 yy = 1  yy −  ( xx +  zz )   xz = 2(1 +  ) xz
E E 0
0 0
2(1 +  ) yz
 zz = 1  zz −  ( xx +  yy )  yz =
E E
Hooke's Law for Uniaxial Stress
If xx = zz = xy = xz = yz = 0

+y  yy
yy  yy =
E
−  yy
xx = zz =
E
+x  xy =  xz =  yz = 0
yy
 Hooke’s Law
Anisotropic materials

• As before, assume stress is linearly related to strain....

• Anisotropic material exhibit two “unusual” features (as
compared to isotropic materials):
– Properties differ with direction (e.g, in general Exx ≠ Eyy ≠ Ezz)
– This can lead to unusual “coupling” effects:
• A normal stress may cause shear strains
• A shear stress may cause normal strains
Hooke’s Law – Anisotropic Materials
Strains caused by xx only:

(What strains are induced by

xx only?)
Hooke’s Law – Anisotropic Materials
Strains caused by xx only:

(What strains are induced by

xx only?)

…for generally anisotropic

materials, xx will induce six
components of strain (i.e.,
xx will induce a 3-D strain
tensor)
Hooke’s Law – Anisotropic Materials
Strains caused by xx only:

xx = k11 xx

yy = k21 xx
zz = k31 xx
xy = k41 xx
yz = k51 xx
zx = k61 xx
Hooke’s Law – Anisotropic Materials
Strains caused by xx only:

xx = k11 xx

yy = k21 xx
zz = k31 xx
xy = k41 xx
yz = k51 xx
zx = k61 xx

In general: k11 ≠ k21 ≠ k31 ≠ k41 ≠ k51 ≠ k61 ≠ 0

Hooke’s Law – Anisotropic Materials
Strains caused by yy only:

xx = k12 yy

yy = k22 yy
zz = k32 yy
xy = k42 yy
yz = k52 yy
zx = k62 yy
Hooke’s Law – Anisotropic Materials
Strains caused by zz only:

xx = k13 zz

yy = k23 zz
zz = k33 zz
xy = k43 zz
yz = k53 zz
zx = k63 zz
Hooke’s Law – Anisotropic Materials
Strains caused by xy only:

xx = k14 xy

yy = k24 xy
zz = k34 xy
xy = k44 xy
yz = k54 xy
zx = k64 xy
Hooke’s Law – Anisotropic Materials
Strains caused by yz only:

xx = k15 yz

yy = k25 yz
zz = k35 yz
xy = k45 yz
yz = k55 yz
zx = k65 yz
Hooke’s Law – Anisotropic Materials
Strains caused by zx only:

xx = k16 zx

yy = k26 zx
zz = k36 zx
xy = k46 zx
yz = k56 zx
zx = k66 zx
Hooke’s Law – Anisotropic Materials
Strain xx caused by all stress components acting
simultaneously:

xx = ?
Hooke’s Law – Anisotropic Materials
Strain xx caused by all stress components acting
simultaneously:

xx = ?

(since strain-stress relation

assumed linear, we can
apply the principle of
superposition and simply
add up contribution of
each stress component):
Hooke’s Law – Anisotropic Materials
Strain xx caused by all stress components acting
simultaneously:

xx = k11 xx

+ k12 yy
+ k13 zz
+ k14 xy
+ k15 yz
+ k16 zx
Hooke’s Law – Anisotropic Materials
Repeating this process for each strain component
results in six simultaneous equations

xx = k11 xx + k12 yy + k13 zz + k14 xy + k15 yz + k16 zx
yy = k21 xx + k22 yy + k23 zz + k24 xy + k25 yz + k26 zx
zz = k31 xx + k32 yy + k33 zz + k34 xy + k35 yz + k36 zx
xy = k41 xx + k42 yy + k43 zz + k44 xy + k45 yz + k46 zx
yz = k51 xx + k52 yy + k53 zz + k54 xy + k55 yz + k56 zx
zx = k61 xx + k62 yy + k63 zz + k64 xy + k65 yz + k66 zx
Hooke’s Law – Anisotropic Materials
…the six eq’s can be expressed using matrix notation

 xx  k11 k12 k13 k14 k15 k16  xx 

  k k 26  
 yy   21 k22 k23 k24 k25  yy 
 zz   k31 k32 k33 k34 k35 k36  zz 
 =   
 xy  k 41 k42 k43 k44 k45 k 46   xy 
 yz  k51 k52 k53 k54 k55 k56  yz 
    
zx  k61 k62 k63 k64 k65 k66    zx 
Hooke’s Law – Anisotropic Materials
Inverting the six equations:

xx = K11 xx + K12 yy + K13 zz + K14 xy + K15 yz + K16 zx
yy = K21 xx + K22 yy + K23 zz + K24 xy + K25 yz + K26 zx
zz = K31 xx + K32 yy + K33 zz + K34 xy + K35 yz + K36 zx
xy = K41 xx + K42 yy + K43 zz + K44 xy + K45 yz + K46 zx
yz = K51 xx + K52 yy + K53 zz + K54 xy + K55 yz + K56 zx
zx = K61 xx + K62 yy + K63 zz + K64 xy + K65 yz + K66 zx

(2.11)
 Hooke’s Law – Anisotropic Materials
Using matrix notation:

 xx   K11 K12 K13 K14 K15 K16  xx 

   K K 26  
 yy   21 K22 K23 K24 K25  yy 
 zz   K 31 K32 K33 K34 K35 K 36  zz 
 =   
  xy   K 41 K42 K43 K44 K45 K 46  xy 
 yz  K51 K52 K53 K54 K55 K 56  yz 
    
zx  K61 K62 K63 K64 K65 K 66   zx 

• Kij = [Kij] = “coefficients of elasticity”

 Hooke’s Law – Anisotropic Materials
Inverting the six equations:

•Apparently, there are 36 coefficients of elasticity, however

• Strain energy considerations show that the [Kij] matrix must be
symmetric….number of independent coefficients reduces from
36 to 21:

 xx  K11 K12 K13 K14 K15 K16  xx 

   K K 26  
 yy   12 K22 K23 K24 K25  yy 
 zz   K13 K23 K33 K34 K35 K 36  zz 
 =   
  xy   K14 K24 K34 K44 K45 K 46  xy 
 yz  K15 K25 K35 K45 K55 K 56  yz 
    
zx  K16 K26 K36 K46 K56 K 66   zx 
 Hooke’s Law – Anisotropic Materials
Anisotropy originates due to microstructure

•Unidirectional composites possess an inherent “principal material

coordinate system”, defined by the fiber orientation
 Hooke’s Law – Anisotropic Materials
Anisotropy originates due to microstructure

•If the stress and strain tensor are described relative to the principal
material coordinate system, then there is no coupling between normal
stress and shear strain, and no coupling between shear stress and
normal strain:

 xx   K11 K12 K13 0 0 0  xx 

   K K 22 K 23 0 0 0  yy 
 yy   12  
 zz   K13 K23 K33 0 0 0  zz 
 =   
  xy   0 0 0 K44 0 0  xy 
 yz   0 0 0 K55 0  yz 
    
zx   K 66   zx 
 Hooke’s Law – Anisotropic Materials
Anisotropy originates due to microstructure

• If fiber distribution
differs in y- and z-
directions, then:

Exx > Eyy ≠ Ezz

 xx   K11 K12 K13 0 0 0  xx 
Orthotropic Material    K K 22 K 23 0 0 0  yy 
 yy   12  
 zz   K13 K23 K33 0 0 0   zz 
 =   
 xy   0 0 0 K44 0 0  xy 
 yz   0 0 0 K55 0  yz 
    
zx   K 66   zx 
(2.21)
 Hooke’s Law – Anisotropic Materials
Anisotropy originates due to microstructure

• If fiber distribution
in y- and z-directions
is identical, then:

Exx > Eyy = Ezz

 xx   K11 K12 K12 0 0 0   
xx
Transversely Isotropic   K12 0 0 
0  yy 
K22 K23
 yy    
Material  zz  K12 K23 K22 0 0 0   zz 
 =  K 22 −K 23  
 xy   0 0 0 0 
0   xy 
 yz   0
2 
0   yz 
  0 0 K 66
  
zx   K 66   zx 


(2.32)
Hooke’s Law
Summary of Key Points

• Hooke’s Law is valid under linear-elastic

conditions only
• The mathematical form of Hooke’s Law depends
on the problem involved:
– Isotropic vs anisotropic materials
– 3-D stress/strains
– Plane stress states
– Plane strain states
– Uniaxial stress
Failure Predictions

• Isotropic metals: methods to predict failure are

reasonably well-developed and are typically based on:
– Yield criterion (i.e., predict the stress or strain tensor
necessary to cause nonlinear behavior), and/or
– Fracture mechanics (i.e., predict the stress or strain tensor
that will cause either stable or unstable crack growth)
• Anisotropic materials (composites): methods to
predict failure are not as well-developed, and often
vary from company-to-company or industry-to-
industry…predicting failure of anisotropic materials is
a active research topic and will not be discussed in this
review
Failure Predictions for Isotropic Metals
Yield criterion
• Context: a structure is subjected to external loads,
causing a 3-D state of stress. What load level will
cause nonlinear behavior (yielding)?

+x

F1
+z xx
+y
F2
M1 xz
zy
xy
zz

F3 yy
F4
M2
 Failure Predictions for Isotropic Metals
Yield criterion
• The von Mises criterion is most commonly used to predict
yielding of isotropic metals such as steel or aluminum alloys
(aka the Maxwell-Huber-Hencky-von Mises criterion)
• Other common yield criterion include the Tresca, Max
Normal Stress, or Mohr’s criterion…these criterion will not
be reviewed here +x

F1
+z xx
+y
F2
M1 xz
zy
xy
zz

F3 yy
F4
M2
 Failure Predictions for Isotropic Metals
The von Mises criterion: 3D stress states
• Yielding occurs if:
1
2

( xx −  yy )2 + ( yy −  zz )2 + ( zz −  xx )2 + 6( xy2 +  yz2 +  xz2 )   Y2 
or, equivalently, if:
1
2

( 11 −  22 )2 + ( 22 −  33 )2 + ( 33 − 11 ) 2   Y2 
where Y = 0.2% offset yield strength (usually tensile)
_
F1 _
F2 xx +x 11 +1
xy
xz
+y yx +2
yz
_ zy 22
F3
zz yy
_ 33
F5 +3
_
F4 +z zx
 Failure Predictions for Isotropic Metals
The von Mises criterion – Plane stress states
• If a plane stress state exists, yielding occurs if:
 2
xx −  xx yy +  yy2 + 3 xy2   Y2 
or, equivalently, if:

 2
11 −  11 22 +  22
2

  Y2
where Y = 0.2% offset yield strength +p +x
xx +1
+x xx +x
11
22
xy xy
+y yx yy yy +y
yy
22
xy +p 11
+y
+z xx +2
 Failure Predictions for Isotropic Metals
Application
• Failure predictions based on yield criterion are typically based on
the concept of “fully plastic” loading
– Elastic-perfectly plastic behavior assumed (aka elastoplastic)
– Failure predicted when entire cross-section is predicted to have yielded
– This approach provides a reasonably conservative estimate, since in reality
metal alloys strain-harden (i.e., are not elastic-perfectly plastic)
 Failure Predictions for Isotropic Metals
Application
• Failure predictions based on yield criterion are typically based on
the concept of “fully plastic” loading
– Elastic-perfectly plastic behavior assumed (aka elastoplastic)
– Failure predicted when entire cross-section is predicted to have yielded
– This approach provides a reasonably conservative estimate, since in reality
metal alloys strain-harden (i.e., are not elastic-perfectly plastic)
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• Inglis (1913) and Westergaard (1939) derived theoretical solutions for stresses
near a crack based on the theory of elasticity (which is, in turn, based on
Hooke’s Law)
• These solutions show that the stresses near a crack tip are “singular”
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• The Inglis and Westergaard solutions are obviously incorrect near the crack
tip…otherwise yy│x→a = ∞ if ∞ ≠ 0.00000….
• Due to very high stresses and stress gradients, traditional yield criterion (e.g., the
von Mises criterion) cannot be applied to predict failure at/near the crack…must
use fracture mechanics instead
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• In 1958 Irwin noted that the general
solutions for stresses near a crack tip +y
can be written in polar coordinates
(and for small r) as:
K
 ij (r, ) = f ij ( ) + higher order terms
2r
+
where:

fij () = dimensionless function Crack

that depends on geometry Crack front

K = the “stress intensity factor,”

whose value depends on geometry,
size and location of the crack,
and the magnitude of loading ; units = stress-length1/2
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• Three types of loading modes are defined:
– Mode I: the “opening mode”… characterized by KI
– Mode II: the “shearing mode”… characterized by KII
– Mode III: the “tearing mode”… characterized by KIII

• Modes I and II are most commonly encountered in practice…if loads associated

with both Mode I and Mode II are present the loading condition is called “mixed
mode” loading
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• Since K increases as loading is increased, it was proposed the failure will occur
when K is increases to a “critical” level.
• The “critical stress intensity factor” for each mode (KIc, KIIc, KIIIc) is treated as a
material property and tabulated like characteristic properties (e.g., E, , or Y)
• Predictions based on Kc are most accurate for brittle materials or materials with
very low levels of ductility (…why?)
• Failure of highly ductile materials is
better predicted using the critical
strain energy release rate (Gc)
…not reviewed here
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• To repeat, the value of the stress intensity factor geometry, size and location of
the crack, and the magnitude of loading
• KI has been tabulated for many geometries (both standard lab specimens and
typical structural geometries)…for example:
– Murakami, Y (ed), Stress Intensity Factors Handbook, Vols 1&2, Pergamon Press,
(1987)
– https://en.wikipedia.org/wiki/Stress_intensity_factor
– (A few are listed in Shukla and Dally, sec 4.4)
• Values for KII and KIII are also available but for fewer geometries
• The tabulated values of KI, KII, KIII are often curve fits of data obtained from:
– Experiments, especially photoelasticity
– Numerical analyses (usually finite-element analyses)
 Failure Predictions for Isotropic Metals
Fracture Mechanics
• For example, KI for an edge crack in a plate under uniaxial stress is (Shukla and
Dally, Sec 4.4, combine Eqs 4.22 and 4.23):
 a a
2
a
3
a 
4
KI = 0 a 1.12 − 0.231  +10.55  − 21.71  + 30.38  
 W  W  W   W  
 Failure Predictions for Isotropic Metals
Fracture Mechanics applied to fatigue failure
• The variation in the stress intensity factor can be used to predict sub-critical
crack growth during fatigue loading:

max,
KIma
Stress () or SIF (KI)

x
min,
KImi
n

Time
 Failure Predictions for Isotropic Metals
Fracture Mechanics applied to fatigue failure
• The variation in the stress intensity factor can be combined with the Paris Law to
predict sub-critical crack growth during fatigue loading:
da
= C(K I ) m
dN

where:

N = load cycle
C, m = experimentally-determined
constants
KI = KImax - KImin

Idealized plot extracted from:

https://en.wikipedia.org/wiki/Paris%27_law