2

Chapter

Stress and Strain: Important

Relationships

Understanding the physical properties of stress and strain is a

prerequisite to utilizing the many methods and results of structural
analysis in design. This chapter provides the denitions and impor-
tant relationships of stress and strain.

2.1 Stress
Stress is simply a distributed force on an external or internal surface
of a body. To obtain a physical feeling of this idea, consider being
submerged in water at a particular depth. The force of the water one
feels at this depth is a pressure, which is a compressive stress, and not
a nite number of concentrated forces. Other types of force distribu-
tions (stress) can occur in a liquid or solid. Tensile (pulling rather than
pushing) and shear (rubbing or sliding) force distributions can also
exist.
Consider a general solid body loaded as shown in Fig. 2.1(a). Pi and
pi are applied concentrated forces and applied surface force distribu-
tions, respectively; and Ri and ri are possible support reaction force
and surface force distributions, respectively. To determine the state of
stress at point Q in the body, it is necessary to expose a surface
containing the point Q. This is done by making a planar slice, or break,
through the body intersecting the point Q. The orientation of this slice
is arbitrary, but it is generally made in a convenient plane where the
state of stress can be determined easily or where certain geometric
relations can be utilized. The rst slice, illustrated in Fig. 2.1(b), is
arbitrarily oriented by the surface normal x. This establishes the yz
plane. The external forces on the remaining body are shown, as well as
the internal force (stress) distribution across the exposed internal

9
10 Formulas for Stress and Strain [CHAP. 2

Figure 2.1

surface containing Q. In the general case, this distribution will not be

uniform along the surface, and will be neither normal nor tangential
to the surface at Q. However, the force distribution at Q will have
components in the normal and tangential directions. These compo-
nents will be tensile or compressive and shear stresses, respectively.
Following a right-handed rectangular coordinate system, the y and z
axes are dened perpendicular to x, and tangential to the surface.
Examine an innitesimal area DAx DyDz surrounding Q, as shown
in Fig. 2.2(a). The equivalent concentrated force due to the force
distribution across this area is DFx , which in general is neither
normal nor tangential to the surface (the subscript x is used to
designate the normal to the area). The force DFx has components in
the x, y, and z directions, which are labeled DFxx , DFxy , and DFxz ,
respectively, as shown in Fig. 2.2(b). Note that the rst subscript
SEC. 2.1] Stress and Strain: Important Relationships 11

Figure 2.2

denotes the direction normal to the surface and the second gives the
actual direction of the force component. The average distributed force
per unit area (average stress) in the x direction isy
DFxx
s xx
DAx

Recalling that stress is actually a point function, we obtain the exact

stress in the x direction at point Q by allowing DAx to approach zero.
Thus,
DFxx
sxx lim
DAx !0 DAx
or,
dFxx
sxx 2:1-1
dAx

Stresses arise from the tangential forces DFxy and DFxz as well, and
since these forces are tangential, the stresses are shear stresses.
Similar to Eq. (2.1-1),
dFxy
txy 2:1-2
dAx
dFxz
txz 2:1-3
dAx

y
Standard engineering practice is to use the Greek symbols s and t for normal (tensile
or compressive) and shear stresses, respectively.
12 Formulas for Stress and Strain [CHAP. 2

Figure 2.3 Stress components.

Since, by denition, s represents a normal stress acting in the same

direction as the corresponding surface normal, double subscripts are
redundant, and standard practice is to drop one of the subscripts and
write sxx as sx . The three stresses existing on the exposed surface at
the point are illustrated together using a single arrow vector for each
stress as shown in Fig. 2.3. However, it is important to realize that the
stress arrow represents a force distribution (stress, force per unit
area), and not a concentrated force. The shear stresses txy and txz are
the components of the net shear stress acting on the surface, where the
net shear stress is given byy
q
tx net t2xy t2xz 2:1-4

To describe the complete state of stress at point Q completely, it

would be necessary to examine other surfaces by making different
planar slices. Since different planar slices would necessitate different
coordinates and different free-body diagrams, the stresses on each
planar surface would be, in general, quite different. As a matter of
fact, in general, an innite variety of conditions of normal and shear
stress exist at a given point within a stressed body. So, it would take an
innitesimal spherical surface surrounding the point Q to understand
and describe the complete state of stress at the point. Fortunately,
through the use of the method of coordinate transformation, it is only
necessary to know the state of stress on three different surfaces to
describe the state of stress on any surface. This method is described in
Sec. 2.3.
The three surfaces are generally selected to be mutually perpendi-
cular, and are illustrated in Fig. 2.4 using the stress subscript notation

y
Stresses can only be added as vectors if they exist on a common surface.
SEC. 2.2] Stress and Strain: Important Relationships 13

Figure 2.4 Stresses on three orthogonal surfaces.

as earlier dened. This state of stress can be written in matrix form,

where the stress matrix s is given by
2 3
sx txy txz
s 4 tyx sy tyz 5 2:1-5
tzx tzy sz

Except for extremely rare cases, it can be shown that adjacent shear
stresses are equal. That is, tyx txy , tzy tyz , and txz tzx , and the
stress matrix is symmetric and written as
2 3
sx txy tzx
s 4 txy sy tyz 5 2:1-6
tzx tyz sz

Plane Stress. There are many practical problems where the stresses
in one direction are zero. This situation is referred to as a case of plane
stress. Arbitrarily selecting the z direction to be stress-free with
sz tyz tyz 0, the last row and column of the stress matrix can
be eliminated, and the stress matrix is written as
 
sx txy
s 2:1-7
txy sy

and the corresponding stress element, viewed three-dimensionally and

down the z axis, is shown in Fig. 2.5.

2.2 Strain and the StressStrain Relations

As with stresses, two types of strains exist: normal and shear strains,
which are denoted by e and g, respectively. Normal strain is the rate of
change of the length of the stressed element in a particular direction.
14 Formulas for Stress and Strain [CHAP. 2

Figure 2.5 Plane stress.

Shear strain is a measure of the distortion of the stressed element, and

has two denitions: the engineering shear strain and the elasticity
shear strain. Here, we will use the former, more popular, denition.
However, a discussion of the relation of the two denitions will be
provided in Sec. 2.4. The engineering shear strain is dened as the
change in the corner angle of the stress cube, in radians.

Normal Strain. Initially, consider only one normal stress sx applied to

the element as shown in Fig. 2.6. We see that the element increases in
length in the x direction and decreases in length in the y and z
directions. The dimensionless rate of increase in length is dened as
the normal strain, where ex , ey , and ez represent the normal strains in

Figure 2.6 Deformation attributed to sx .

SEC. 2.2] Stress and Strain: Important Relationships 15

the x, y, and z directions respectively. Thus, the new length in any

direction is equal to its original length plus the rate of increase
(normal strain) times its original length. That is,

Dx0 Dx ex Dx; Dy0 Dy ey Dy; Dz0 Dz ez Dz 2:2-1

There is a direct relationship between strain and stress. Hookes law

for a linear, homogeneous, isotropic material is simply that the normal
strain is directly proportional to the normal stress, and is given by

1
ex s  nsy sz  2:2-2a
E x
1
ey sy  nsz sx  2:2-2b
E
1
ez sz  nsx sy  2:2-2c
E

where the material constants, E and n, are the modulus of elasticity

(also referred to as Youngs modulus) and Poissons ratio, respectively.
Typical values of E and n for some materials are given in Table 2.1 at
the end of this chapter.
If the strains in Eqs. (2.2-2) are known, the stresses can be solved for
simultaneously to obtain

E
sx 1  nex ney ez  2:2-3a
1 n1  2n
E
sy 1  ney nez ex  2:2-3b
1 n1  2n
E
sz 1  nez nex ey  2:2-3c
1 n1  2n

For plane stress, with sz 0, Eqs. (2.2-2) and (2.2-3) become

1
ex s  nsy 2:2-4a
E x
1
ey sy  nsx 2:2-4b
E
n
ez  sx sy 2:2-4c
E
16 Formulas for Stress and Strain [CHAP. 2

and

E
sx e ney 2:2-5a
1  n2 x
E
sy e nex 2:2-5b
1  n2 y

Shear Strain. The change in shape of the element caused by the shear
stresses can be rst illustrated by examining the effect of txy alone as
shown in Fig. 2.7. The engineering shear strain gxy is a measure of the
skewing of the stressed element from a rectangular parallelepiped. In
Fig. 2.7(b), the shear strain is dened as the change in the angle BAD.
That is,

gxy BAD  B0 A0 D 0

where gxy is in dimensionless radians.

For a linear, homogeneous, isotropic material, the shear strains in
the xy, yz, and zx planes are directly related to the shear stresses by

txy
gxy 2:2-6a
G
tyz
gyz 2:2-6b
G
tzx
gzx 2:2-6c
G

where the material constant, G, is called the shear modulus.

Figure 2.7 Shear deformation.

SEC. 2.3] Stress and Strain: Important Relationships 17

It can be shown that for a linear, homogeneous, isotropic material

the shear modulus is related to Poissons ratio by (Ref. 1)

E
G 2:2-7
21 n

2.3 Stress Transformations

As was stated in Sec. 2.1, knowing the state of stress on three
mutually orthogonal surfaces at a point in a structure is sufcient to
generate the state of stress for any surface at the point. This is
accomplished through the use of coordinate transformations. The
development of the transformation equations is quite lengthy and is
not provided here (see Ref. 1). Consider the element shown in Fig.
2.8(a), where the stresses on surfaces with normals in the x, y, and z
directions are known and are represented by the stress matrix
2 3
sx txy tzx
sxyz 4 txy sy tyz 5 2:3-1
tzx tyz sz

Now consider the element, shown in Fig. 2.8(b), to correspond to the

state of stress at the same point but dened relative to a different set of
surfaces with normals in the x0 , y0 , and z0 directions. The stress matrix
corresponding to this element is given by
2 3
sx0 tx0 y0 tz0 x0
sx0 y0 z0 4 tx0 y0 sy0 ty0 z0 5 2:3-2
tz0 x0 ty0 z0 sz0

To determine sx0 y0 z0 by coordinate transformation, we need to

establish the relationship between the x0 y0 z0 and the xyz coordinate
systems. This is normally done using directional cosines. First, let us
consider the relationship between the x0 axis and the xyz coordinate
system. The orientation of the x0 axis can be established by the angles
yx0 x , yx0 y , and yx0 z , as shown in Fig. 2.9. The directional cosines for x0 are
given by

lx0 cos yx0 x ; mx0 cos yx0 y ; nx0 cos yx0 z 2:3-3

Similarly, the y0 and z0 axes can be dened by the angles yy0 x , yy0 y , yy0 z
and yz0 x , yz0 y , yz0 z , respectively, with corresponding directional cosines

ly0 cos yy0 x ; my0 cos yy0 y ; ny0 cos yy0 z 2:3-4
lz0 cos yz0 x ; mz0 cos yz0 y ; nz0 cos yz0 z 2:3-5
18 Formulas for Stress and Strain [CHAP. 2

Figure 2.8 The stress at a point using different coordinate systems.

It can be shown that the transformation matrix

2 3
lx0 mx0 nx 0
T 4 ly0 my0 ny 0 5 2:3-6
lz0 mz0 nz0
SEC. 2.3] Stress and Strain: Important Relationships 19

Figure 2.9 Coordinate transformation.

transforms a vector given in xyz coordinates, fVgxyz , to a vector in x0 y0 z0

coordinates, fVgx0 y0 z0 , by the matrix multiplication

fVgx0 y0 z0 TfVgxyz 2:3-7

Furthermore, it can be shown that the transformation equation for the

stress matrix is given by (see Ref. 1)

sx0 y0 z0 Tsxyz TT 2:3-8

where TT is the transpose of the transformation matrix T, which is

simply an interchange of rows and columns. That is,

2 3
lx0 ly0 lz0
TT 4 mx0 my0 mz 0 5 2:3-9
nx 0 ny0 nz 0

The stress transformation by Eq. (2.3-8) can be implemented very

easily using a computer spreadsheet or mathematical software. Den-
ing the directional cosines is another matter. One method is to dene
the x0 y0 z0 coordinate system by a series of two-dimensional rotations
from the initial xyz coordinate system. Table 2.2 at the end of this
chapter gives transformation matrices for this.
20 Formulas for Stress and Strain [CHAP. 2

EXAMPLE
The state of stress at a point relative to an xyz coordinate system is given by
the stress matrix
2 3
8 6 2
sxyz 4 6 4 25 MPa
2 2 5

Determine the state of stress on an element that is oriented by rst rotating

the xyz axes 45 about the z axis, and then rotating the resulting axes 30
about the new x axis.
Solution. The surface normals can be found by a series of coordinate
transformations for each rotation. From Fig. 2.10(a), the vector components
for the rst rotation can be represented by
8 9 2 38 9
< x1 = cos y sin y 0 < x =
y 4  sin y cos y 0 5 y a
: 1; : ;
z1 0 0 1 z

The last rotation establishes the x0 y0 z0 coordinates as shown in Fig. 2.10(b), and
they are related to the x1 y1 z1 coordinates by
8 09 2 38 9
<x = 1 0 0 < x1 =
y0 4 0 cos j sin j 5 y1 b
: 0; : ;
z 0  sin j cos j z1

Substituting Eq. (a) in (b) gives

8 09 2 32 38 9
>
<x > = 1 0 0 cos y sin y 0 > <x> =
0 6 76 7
y 40 cos j sin j 54  sin y cos y 0 5 y
>
: 0> ; >
: > ;
z 0  sin j cos j 0 0 1 z
2 38 9
cos y sin y 0 <x>
> =
6 7
4  sin y cos j cos y cos j sin j 5 y c
: >
> ;
sin y sin j  cos y sin j cos j z

Figure 2.10
SEC. 2.3] Stress and Strain: Important Relationships 21

Equation (c) is of the form of Eq. (2.3-7). Thus, the transformation matrix is
2 3
cos y sin y 0
T 4  sin y cos j cos y cos j sin j 5 d
sin y sin j  cos y sin j cos j

Substituting y 45 and j 30 gives

2 p p 3
2 2 2 2 0
1 6 p p 7
T 4  6 6 2 5 e
4 p p p
2  2 2 3

The transpose of T is
2 p p p 3
2 2  6 2
16 p p p 7
TT 4 2 2 6  25 f
4 p
0 2 2 3

From Eq. (2.3-8),

2 p p 32 3 2 p p p 3
2 2 2 2 0 8 6 2 2 2  6 2
1 6 p p 76 7 1 6 p p p 7
sx0 y0 z0 4 6 6 2 54 6 4 25 42 2 6  25
4 p p p 4 p
2  2 2 3 2 2 5 0 2 2 3

This matrix multiplication can be performed simply using either a computer

spreadsheet or mathematical software, resulting in
2 3
4 5:196 3
sx0 y0 z0 4 5:196 4:801 2:714 5 MPa
3 2:714 8:199

Stresses on a Single Surface. If one was concerned about the state of

stress on one particular surface, a complete stress transformation
would be unnecessary. Let the directional cosines for the normal of
the surface be given by l, m, and n. It can be shown that the normal
stress on the surface is given by

s sx l2 sy m2 sz n2 2txy lm 2tyz mn 2tzx nl 2:3-10

and the net shear stress on the surface is

t sx l txy m tzx n2 txy l sy m tyz n2

tzx l tyz m sz n2  s2 1=2 2:3-11
22 Formulas for Stress and Strain [CHAP. 2

The direction of t is established by the directional cosines

1
lt sx  sl txy m tzx n
t
1
mt txy l sy  sm tyz n 2:3-12
t
1
nt tzx l tyz m sz  sn
t

EXAMPLE
The state of stress at a particular point relative to the xyz coordinate system is
2 3
14 7 7
sxyz 4 7 10 05 kpsi
7 0 35

Determine the normal and shear stress on a surface at the point where the
surface is parallel to the plane given by the equation

2x  y 3z 9

Solution. The normal to the surface is established by the directional

numbers of the plane and are simply the coefcients of x, y, and z terms of
the equation of the plane. Thus, the directional numbers are 2, 1, and 3. The
directional cosines of the normal to the surface are simply the normalized
values of the directional numbers, which are the directional numbers divided
q p
by 22 12 32 14. Thus

p p p
l 2= 14; m 1= 14; n 3= 14

From the stress matrix, sx 14, txy 7, tzx 7, sy 10, tyz 0, and sz
35 kpsi. Substituting the stresses and directional cosines into Eq. (2.3-10)
gives
p p p p p
s 142= 142 101= 142 353= 142 272= 141= 14
p p p p
201= 143= 14 273= 142= 14 19:21 kpsi

The shear stress is determined from Eq. (2.3-11), and is

p p p
t f142= 14 71= 14 73= 142
p p p
72= 14 101= 14 03= 142
p p p
72= 14 01= 14 353= 142  19:212 g1=2 14:95 kpsi
SEC. 2.3] Stress and Strain: Important Relationships 23

From Eq. (2.3-12), the directional cosines for the direction of t are

1 p p p
lt 14  19:212= 14 71= 14 73= 14 0:687
14:95
1 p p p
mt 72= 14 10  19:211= 14 03= 14 0:415
14:95
1 p p p
nt 72= 14 01= 14 35  19:213= 14 0:596
14:95

Plane Stress. For the state of plane stress shown in Fig. 2.11(a),
sz tyz tzx 0. Plane stress transformations are normally per-
formed in the xy plane, as shown in Fig. 2.11(b). The angles relating
the x0 y0 z0 axes to the xyz axes are

yx0 x y; yx0 y 90  y; yx0 z 90

yy0 x y 90 ; yy0 y y; yy0 z 90
yz0 x 90 ; yz0 y 90 ; yz0 z 0

Thus the directional cosines are

lx0 cos y mx0 sin y nx 0 0

ly0  sin y my0 cos y ny 0 0
lz0 0 mz0 0 nz 0 1

The last rows and columns of the stress matrices are zero so the
stress matrices can be written as
 
sx txy
sxy 2:3-13
txy sy

Figure 2.11 Plane stress transformations.

24 Formulas for Stress and Strain [CHAP. 2

and
 
sx 0 tx0 y0
sx0 y0 2:3-14
tx0 y0 sy0

Since the plane stress matrices are 2  2, the transformation matrix

and its transpose are written as
   
cos y sin y T cos y  sin y
T ; T 2:3-15
 sin y cos y sin y cos y

Equations (2.3-13)(2.3-15) can then be substituted into Eq. (2.3-8) to

perform the desired transformation. The results, written in long-hand
form, would be

sx0 sx cos2 y sy sin2 y 2txy cos y sin y

sy0 sx sin2 y sy cos2 y  2txy cos y sin y 2:3-16
2
tx0 y0 sx  sy sin y cos y txy cos y  sin y
2

If the state of stress is desired on a single surface with a normal

rotated y counterclockwise from the x axis, the rst and third equa-
tions of Eqs. (2.3-16) can be used as given. However, using trigono-
metric identities, the equations can be written in slightly different
form. Letting s and t represent the desired normal and shear stresses
on the surface, the equations are

sx sy sx  sy
s cos 2y txy sin 2y
2 2
2:3-17
sx  sy
t sin 2y txy cos 2y
2

Equations (2.3-17) represent a set of parametric equations of a circle in

the st plane. This circle is commonly referred to as Mohrs circle and is
generally discussed in standard mechanics of materials textbooks.
This serves primarily as a teaching tool and adds little to applications,
so it will not be represented here (see Ref. 1).

Principal Stresses. In general, maximum and minimum values of the

normal stresses occur on surfaces where the shear stresses are zero.
These stresses, which are actually the eigenvalues of the stress
matrix, are called the principal stresses. Three principal stresses
exist, s1 , s2 , and s3 , where they are commonly ordered as s1 5 s2 5 s3 .
SEC. 2.3] Stress and Strain: Important Relationships 25

Considering the stress state given by the matrix of Eq. (2.3-1) to be

known, the principal stresses sp are related to the given stresses by

sx  sp lp txy mp tzx np 0
txy lp sy  sp mp tyz np 0 2:3-18
tzx lp tyz mp sz  sp np 0

where lp , mp , and np are the directional cosines of the normals to the

surfaces containing the principal stresses. One possible solution to
Eqs. (2.3-18) is lp mp np 0. However, this cannot occur, since

l2p m2p n2p 1 2:3-19

To avoid the zero solution of the directional cosines of Eqs. (2.3-18), the
determinant of the coefcients of lp , mp , and np in the equation is set to
zero. This makes the solution of the directional cosines indeterminate
from Eqs. (2.3-18). Thus,
 
 sx  sp txy tzx 
 
 txy s  s t 0
 y p yz 
 tzx tyz sz  sp 

Expanding the determinant yields

s3p  sx sy sz s2p sx sy sy sz sz sx  t2xy  t2yz  t2zx sp

 sx sy sz 2txy tyz tzx  sx t2yz  sy t2zx  sz t2xy 0 2:3-20

where Eq. (2.3-20) is a cubic equation yielding the three principal

stresses s1 , s2 , and s3 .
To determine the directional cosines for a specic principal stress,
the stress is substituted into Eqs. (2.3-18). The three resulting equa-
tions in the unknowns lp , mp , and np will not be independent since
they were used to obtain the principal stress. Thus, only two of Eqs.
(2.3-18) can be used. However, the second-order Eq. (2.3-19) can be
used as the third equation for the three directional cosines. Instead of
solving one second-order and two linear equations simultaneously, a
simplied method is demonstrated in the following example.y

y
Mathematical software packages can be used quite easily to extract the eigenvalues
(sp ) and the corresponding eigenvectors (lp , mp , and np ) of a stress matrix. The reader is
urged to explore software such as Mathcad, Matlab, Maple, and Mathematica, etc.
26 Formulas for Stress and Strain [CHAP. 2

EXAMPLE
For the following stress matrix, determine the principal stresses and the
directional cosines associated with the normals to the surfaces of each
principal stress.

2 3
3 1 1
s 4 1 0 25 MPa
1 2 0

Solution. Substituting sx 3, txy 1, tzx 1, sy 0, tyz 2, and sz 0 into

Eq. (2.3-20) gives

s2p  3 0 0s2p 30 00 03  22  12  12 sp

 300 2211  322  012  012  0

which simplies to

s2p  3s2p  6sp 8 0 a

The solutions to the cubic equation are sp 4, 1, and 2 MPa. Following the
conventional ordering,

s1 4 MPa; s2 1 MPa; s3 2 MPa

The directional cosines associated with each principal stress are determined
independently. First, consider s1 and substitute sp 4 MPa into Eqs. (2.3-18).
This results in

l1 m1 n1 0 b
l1  4m1 2n1 0 c
l1 2m1  4n1 0 d

where the subscript agrees with that of s1 .

Equations (b), (c), and (d) are no longer independent since they were used to
determine the values of sp . Only two independent equations can be used, and
in this example, any two of the above can be used. Consider Eqs. (b) and (c),
which are independent. A third equation comes from Eq. (2.3-19), which is
nonlinear in l1 , m1 , and n1 . Rather than solving the three equations simulta-
neously, consider the following approach.
Arbitrarily, let l1 1 in Eqs. (b) and (c). Rearranging gives

m1 n1 1
4m1  2n1 1
SEC. 2.3] Stress and Strain: Important Relationships 27

solving these simultaneously gives m1 n1 12. These values of l1 , m1 , and n1

do not satisfy Eq. (2.3-19). However, all that remains is to normalize their
q p
values by dividing by 12 122 122 6=2: Thus,y
p p
l1 12= 6 6=3
p p
m1 1=22= 6 6=6
p p
n1 1=22= 6 6=6

Repeating the same procedure for s2 1 MPa results in

p p p
l2 3=3; m2  3=3; n2  3=3

and for s3 2 MPa

p p
l3 0; m3 2=2; n3  2=2

If two of the principal stresses are equal, there will exist an innite set
of surfaces containing these principal stresses, where the normals of
these surfaces are perpendicular to the direction of the third principal
stress. If all three principal stresses are equal, a hydrostatic state of
stress exists, and regardless of orientation, all surfaces contain the
same principal stress with no shear stress.

Principal Stresses, Plane Stress. Considering the stress element shown

in Fig. 2.11(a), the shear stresses on the surface with a normal in the z
direction are zero. Thus, the normal stress sz 0 is a principal stress.
The directions of the remaining two principal stresses will be in the xy
plane. If tx0 y0 0 in Fig. 2.11(b), then sx0 would be a principal stress, sp
with lp cos y, mp sin y, and np 0. For this case, only the rst two
of Eqs. (2.3-18) apply, and are

sx  sp cos y txy sin y 0

2:3-21
txy cos y sy  sp sin y 0

As before, we eliminate the trivial solution of Eqs. (2.3-21) by setting

the determinant of the coefcients of the directional cosines to zero.
That is,
 
 s  s txy 
 x p 
  sx  sp sy  sp  t2xy
 txy sy  sp 
s2p  sx sy sp sx sy  t2xy 0 2:3-22

y
This method has one potential aw. If l1 is actually zero, then a solution would not
result. If this happens, simply repeat the approach letting either m1 or n1 equal unity.
28 Formulas for Stress and Strain [CHAP. 2

Equation (2.3-22) is a quadratic equation in sp for which the two

solutions are
 q
1
sp sx sy  sx  sy 2 4t2xy 2:3-23
2

Since for plane stress, one of the principal stresses (sz ) is always zero,
numbering of the stresses (s1 5 s2 5 s3 ) cannot be performed until Eq.
(2.3-23) is solved.
Each solution of Eq. (2.3-23) can then be substituted into one of Eqs.
(2.3-21) to determine the direction of the principal stress. Note that if
sx sy and txy 0, then sx and sy are principal stresses and Eqs.
(2.3-21) are satised for all values of y. This means that all stresses in
the plane of analysis are equal and the state of stress at the point is
isotropic in the plane.

EXAMPLE
Determine the principal stresses for a case of plane stress given by the stress
matrix
 
5 4
s kpsi
4 11

Show the element containing the principal stresses properly oriented with
respect to the initial xyz coordinate system.
Solution. From the stress matrix, sx 5, sy 11, and txy 4 kpsi and Eq.
(2.3-23) gives
 q
sp 12 5 11  5  112 442 13; 3 kpsi

Thus, the three principal stresses (s1 ; s2 ; s3 ), are (13, 3, 0) kpsi, respectively.
For directions, rst substitute s1 13 kpsi into either one of Eqs. (2.3-21).
Using the rst equation with y y1

sx  s1 cos y1 txy sin y1 5  13 cos y1 4 sin y1 0

or
 
8
y1 tan1  63:4
4

Now for the other principal stress, s2 3 kpsi, the rst of Eqs. (2.3-21) gives

sx  s2 cos y2 txy sin y2 5  3 cos y2 4 sin y2 0

or
 
2
y2 tan1 26:6
4
SEC. 2.3] Stress and Strain: Important Relationships 29

Figure 2.12 Plane stress example.

Figure 2.12(a) illustrates the initial state of stress, whereas the orientation
of the element containing the in-plane principal stresses is shown in Fig.
2.12(b).

Maximum Shear Stresses. Consider that the principal stresses for a

general stress state have been determined using the methods just
described and are illustrated by Fig. 2.13. The 123 axes represent the
normals for the principal surfaces with directional cosines determined
by Eqs. (2.3-18) and (2.3-19). Viewing down a principal stress axis
(e.g., the 3 axis) and performing a plane stress transformation in the
plane normal to that axis (e.g., the 12 plane), one would nd that the
shear stress is a maximum on surfaces 45 from the two principal
stresses in that plane (e.g., s1 , s2 ). On these surfaces, the maximum
shear stress would be one-half the difference of the principal stresses
[e.g., tmax s1  s2 =2] and will also have a normal stress equal to the
average of the principal stresses [e.g., save s1 s2 =2]. Viewing
along the three principal axes would result in three shear stress

Figure 2.13 Principal stress state.

30 Formulas for Stress and Strain [CHAP. 2

maxima, sometimes referred to as the principal shear stresses. These

stresses together with their accompanying normal stresses are

Plane 1; 2: tmax 1;2 s1  s2 =2; save 1;2 s1 s2 =2

Plane 2; 3: tmax 2;3 s2  s3 =2; save 2;3 s2 s3 =2
Plane 1; 3: tmax 1;3 s1  s3 =2; save 1;3 s1 s3 =2
2:3-24

Since conventional practice is to order the principal stresses by

s1 5 s2 5 s3 , the largest shear stress of all is given by the third of
Eqs. (2.3-24) and will be repeated here for emphasis:

tmax s1  s3 =2 2:3-25

EXAMPLE
In the previous example, the principal stresses for the stress matrix
 
5 4
s kpsi
4 11

were found to be (s1 ; s2 ; s3 13; 3; 0 kpsi. The orientation of the element

containing the principal stresses was shown in Fig. 2.12(b), where axis 3 was
the z axis and normal to the page. Determine the maximum shear stress and
show the orientation and complete state of stress of the element containing
this stress.
Solution. The initial element and the transformed element containing the
principal stresses are repeated in Fig. 2.14(a) and (b), respectively. The
maximum shear stress will exist in the 1, 3 plane and is determined by
substituting s1 13 and s3 0 into Eqs. (2.3-24). This results in

tmax 1;3 13  0=2 6:5 kpsi; save 1;3 13 0=2 6:5 kpsi

To establish the orientation of these stresses, view the element along the axis
containing s2 3 kpsi [view A, Fig. 2.14(c)] and rotate the surfaces 45 as
shown in Fig. 2.14(c).
The directional cosines associated with the surfaces are found through
successive rotations. Rotating the xyz axes to the 123 axes yields
8 9 2 8 9
>
<1> = cos 63:4  sin 63:4 0>
<x>=
6
2 4 sin 63:4 cos 63:4 0 y
>
: > ; >
: >;
3 0 0 1 z
2 38 9
0:4472 0:8944 0 ><x> =
6 7
4 0:8944 0:4472 0 5 y a
>
: > ;
0 0 1 z
SEC. 2.3] Stress and Strain: Important Relationships 31

Figure 2.14 Plane stress maximum shear stress.

A counterclockwise rotation of 45 of the normal in the 3 direction about axis 2

is represented by
8 09 2 38 9
>
<x >= cos 45 0  sin 45 > <1> =
6 7
y0 4 0 1 0 5 2
>
: 0>; >
: > ;
z sin 45 0 cos 45 3
2 38 9
0:7071 0 0:7071 ><1> =
6 7
4 0 1 0 5 2 b
>
: ; >
0:7071 0 0:7071 3

Thus,
8 09 2 32 38 9
>
<x > = 0:7071 0 0:7071 0:4472 0:8944 0 <>x> =
0 6 76 7
y 4 0 1 0 54 0:8944 0:4472 0 5 y
>
: 0> ; >
: > ;
z 0:7071 0 0:7071 0 0 1 z
2 38 9
0:3162 0:6325 0:7071 > <x> =
6 7
4 0:8944 0:4472 0 5 y
> >
: ;
0:3162 0:6325 0:7071 z
32 Formulas for Stress and Strain [CHAP. 2

The directional cosines for Eq. (2.1-14c) are therefore

2 3 2 3
nx0 x nx0 y nx0 z 0:3162 0:6325 0:7071
4 ny0 x ny0 y ny0 z 5 4 0:8944 0:4472 0 5
nz0 x nz0 y nz0 z 0:3162 0:6325 0:7071

The other surface containing the maximum shear stress can be found similarly
except for a clockwise rotation of 45 for the second rotation.

2.4 Strain Transformations

The equations for strain transformations are identical to those for
stress transformations. However, the engineering strains as dened in
Sec. 2.2 will not transform. Transformations can be performed if the
shear strain is modied. All of the equations for the stress transforma-
tions can be employed simply by replacing s and t in the equations by e
and g=2 (using the same subscripts), respectively. Thus, for example,
the equations for plane stress, Eqs. (2.3-16), can be written for strain
as

ex0 ex cos2 y ey sin2 y gxy cos y sin y

ey0 ex sin2 y ey cos2 y  gxy cos y sin y 2:4-1
2
gx0 y0 2ex  ey sin y cos y gxy cos y  sin y
2

2.5 Reference
1. Budynas, R. G.: Advanced Strength and Applied Stress Analysis, 2nd ed., McGraw-
Hill, 1999.
SEC. 2.6] Stress and Strain: Important Relationships 33

2.6 Tables
TABLE 2.1 Material propertiesy

Thermal
Modulus of expansion
elasticity, E coefcient, a
Poissons
Material Mpsi GPa ratio, n m= F m= C
Aluminum alloys 10.5 72 0.33 13.1 23.5
Brass (65=35) 16 110 0.32 11.6 20.9
Concrete 4 34 0.20 5.5 9.9
Copper 17 118 0.33 9.4 16.9
Glass 10 69 0.24 5.1 9.2
Iron (gray cast) 13 90 0.26 6.7 12.1
Steel (structural) 29.5 207 0.29 6.5 11.7
Steel (stainless) 28 193 0.30 9.6 17.3
Titanium (6 A1=4 V) 16.5 115 0.34 5.2 9.5
y
The values given in this table are to be treated as approximations of the true behavior of an
actual batch of the given material.

TABLE 2.2 Transformation matrices for positive

rotations about an axisy

Axis Transformation matrix

x axis:
8 9 2 38 9
< x1 = 1 0 0 <x=
y1 4 0 cos y sin y 5 y
: ; : ;
z1 0  sin y cos y z

y axis:

8 9 2 38 9
< x1 = cos y 0  sin y < x =
y1 4 0 1 0 5 y
: ; : ;
z1 sin y 0 cos y z

z axis:

8 9 2 38 9
< x1 = cos y sin y 0 < x =
y1 4  sin y cos y 0 5 y
: ; : ;
z1 0 0 1 z

y
A positive rotation about a given axis is counterclockwise about
the axis (as viewed from the positive axis direction).
34 Formulas for Stress and Strain [CHAP. 2

TABLE 2.3 Transformation equations

General state of stress

sx0 y0 z0 Tsxyz TT

where
2 3 2 3 2 3
sx0 tx0 y0 tz0 x0 lx0 mx0 nx0 sx txy tzx
6 ty0 z0 7 6 7 6 tyz 7
sx0 y0 z0 4 tx0 y0 sy0 5; T 4 ly0 my0 ny0 5; sxyz 4 txy sy 5
tz0 x0 ty0 z0 sz0 lz0 mz0 nz0 tzx tyz sz

Stresses on a single surface (l, m, n are directional cosines of surface normal)

s sx l2 sy m2 sz n2 2txy lm 2tyz mn 2tzx nl
t sx l txy m tzx n2 txy l sy m tyz n2 tzx l tyz m sz n2  s2 1=2
1
lt sx  sl txy m tzx n
t
1
mt txy l sy  sm tyz n
t
1
nt tzx l tyz m sz  sn
t
lt , mt , and nt are directional cosines for the direction of t.
Plane stress (y is counterclockwise from x axis to surface normal, x0 )

s 12 sx sy 12 sx  sy cos 2y txy sin 2y

t  12 sx  sy sin 2y txy cos 2y

Principal stresses (general case)

s3p  sx sy sz s2p sx sy sy sz sz sx  t2xy  t2yz  t2zx sp

 sx sy sz 2txy tyz tzx  sx t2yz  sy t2zx  sz t2xy 0
Directional cosines (lp , mp , np ) are found from three of the following equations:
9
sx  sp lp txy mp tzx np 0 =
txy lp sy  sp mp tyz np 0 select two independent equations
;
tzx lp tyz mp sz  sp np 0

l2p m2p n2p 1

Principal stresses (plane stress) One principal stress is zero and the remaining two are
given by
 q
sp 12 sx sy  sx  sy 2 4t2xy
Angle of surface normal relative to the x axis is given by
!
1 sp  sx
yp tan
txy