4.

SHAFTS
A shaft is an element used to transmit power and torque, and it can support
reverse bending (fatigue). Most shafts have circular cross sections, either solid or tubular.
The difference between a shaft and an axle is that the shaft rotates to transmit power, and
that it is subjected to fatigue. An axle is just like a round cantilever beam, so it is not
subjected to fatigue.

Shafts have different means to transmit power and torque. For example, it can use
gears, sprockets, pulleys, etc., and also have some grooves to keep these elements rigid
and avoid their vibration, such as key seats, retaining ring grooves, etc. Also, to be able to
avoid vibration of the elements, and assure an efficient transmission of power and torque,
some changes in the cross-section of the shaft can be made.

All these elements, because of their weight, of the torque they are transmitting,
and of the grooves made to avoid excessive axial movement, produce stresses, and as a
consequence, strains. It has been demonstrated that stresses and strains are directly
related if they are in the elastic range (by Hookes law), and that if a strain is produced,
then as a consequence a stress has to be produced, and vice versa. Also, there are four
types of shafts, and each one of them has a different way to calculate their corresponding
stresses. In this chapter, there will be analyzed four cases: uniform shaft, multiple-section
shaft, variable-cross-section shaft and relative rotation. All of these shafts are subjected to
torsion, so we will review first the concepts of it.

Torsion refers to the twisting of a straight bar when it is loaded by moments (or
torques) that tends to produce rotation about the longitudinal axis of the bar. (Gere &
Timoshenko, 1997). As demonstrated in the Chapter 3, torsion produces an angle of twist
in the shaft, and this angle of twist produces a strain, which varies linearly with the
distance from the center of the shaft to the surface. As a consequence of the strain
produced by the torsion, also some stresses will be produced, which are shearing stresses.
If we are working in the elastic range, then these shear stresses can be determined from
Hookes law in shear:
= G
where is the shear strain, and G is the modulus of rigidity (or shear modulus of
elasticity). If we combine this equation with the equation to calculate the shear strain we
will get that
max max

r
G r G = = =
1
where
max
is the maximum shearing stress, which occurs at the surface of the shaft
(where =
max
) as seen in Figure 14.


1
From Gere & Timoshenko, Mechanics of Materials, PWS Publishing Company, United States, Fourth
Edition, 1997, p. 192
Figure 14 15 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

What is seen in Figure 14 is how stresses act on the plane of the cross-section, but
this is not the only plane where they act. They are accompanied by shearing stresses of
the same magnitude, but these ones act on a longitudinal plane. This comes from the fact
that equal shear stresses always exist on mutually perpendicular planes (Gere &
Timoshenko, 19797).

Now, we have to take a look to the torsion formula, which states the relationship
between the torque and the shearing stresses. These stresses are acting continuously on
the cross-section of the shaft, and because of this they produce a resultant in the form of a
moment. This moment is equal to the torque applied to the shaft. Considering Figure 15
(which represents a cut in the middle of a bar with an applied torque), in can be seen that
shearing forces are produced, and that they can be calculated at any distance from the
axis. If we make the sum of the moments around the axis, then we will get
that ,

= dF T

Figure 15 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

and as the force is related to the shearing stress by dF = dA, then the torque will
become

= ) ( dA T
2
. Finally, the shearing stress can be obtained from
p
I
T

= , where
T is the applied torque, is the distance from the axis to the point where the shearing
stress is being calculated, and I
p
is the polar moment of inertia (
32
4
d
I
p

=

). If the shaft
is hollow, then the formula used is the same:
J
T
p

= where J is the polar second
moment of area (which is the same as the polar moment of inertia), but it is calculated
as ) (
32
4 4
0 i
D D J =

.
An important thing for the designer is to identify the direction of the shearing
stresses in the shafts produced by the applied torque. Now it is known that to calculate
the magnitude of the shearing stresses at any point from the axis to the surface, the
formula
J
T
p

= can be used. To establish the direction of them, a cut should be made
in the middle of the shaft and it must be perpendicular to the longitudinal axis. If the two
parts of the shafts are now separated, each one must have the same shearing stresses in
that plane, but in opposite direction (see Figure 16). This is because of equilibrium. At
the beginning, the direction of the shearing stress is assumed. By taking moments again
about the axis of the shaft, the shearing stress must be balancing the applied torque. Once
the static equilibrium is established, the direction of the shearing stress is known. The

2
From Beer and Johnston et. al., Interactive Tutorial, Chapter 3, 2001
stresses acting on the surface of a shaft are as shown in Figure 17. They act following the
path of the perimeter of the circle in one direction. If the torque is applied on the opposite
direction,

Figure 16 (Beer & J ohnston et. al. Interactive Tutorial, 2001)
then these stresses will be acting also in the opposite direction, following the same
principle stated before to establish the static equilibrium between torque and stresses. The
complete description of how stresses are behaving in some points of the shaft is seen in
figure 17. At each point, there are two stress free surfaces, and the other four surfaces
have stresses acting as shown.

Figure 17 (Beer & J ohnston et. al. Interactive Tutorial, 2001)
If we have a stepped circular shaft with some torques applied, then the principle
of static equilibrium must be accomplished too, and this way we will be able to find the
magnitude and the direction of the inner torque at each section, and with this we will be
capable of calculating the shearing stresses in that particular section. For example, in
Figure 18a it is seen that we have a shaft with four different torques applied, and that it is
cut on one section of the shaft (where the plane of section 2 is) (Figure 18b). By applying
the

(a)

(b)
Figure 18 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

static equilibrium principle, then we would have that the torque 2 (T
2
) is equal to the
summation of the other two torques:
C D
T T T =
2
. Once this torque T
2
is known the
stresses in that section can be calculated. It is important to understand that the torque T
2

is only valid for that section, and that it describes only the inner torque for that section. If
we want to calculate the inner torque of section 1 (the section on the left before the step),
then we would have to establish again the static equilibrium equation, and calculate the
corresponding torque to calculate the corresponding stresses produced by it.

So far, we have seen that a normal and shearing stresses combination can occur
under the same loading conditions, and they will be oriented at some angle from the axis
of the shaft. If an element oriented at 45 to the axis of the shaft is considered (Figure
19a), then the normal and shearing stresses will be on inclined planes. To determine the
magnitude of the stresses, we must consider two triangular elements (Figure 19b). As we
can see, the vectors resulting from forces AB, BC and CD are the maximum shearing
stresses (
max
). The normal stress is the one which is expressed as F, and as seen it must
be

(a)


(b)
Figure 19 (Beer & J ohnston et. al. Interactive Tutorial, 2001)
perpendicular to the surface. To determine the magnitude of the forces producing the
normal stress, we have to use free body diagrams
3
, from where we would find that

3
To see the demonstration, take a look at: Beer & J ohnston et. al., Mechanics of Materials, McGraw-Hill,
New York, Third Edition, 2001, p. 142
2
0 max
A F =
2 '
0 max
A F =
To obtain the magnitude of the stresses, we just have to divide this force over the cross-
sectional area of each face. From here we can conclude that the normal stress at 45
is
max 45
=

. If the element is not inclined, then it will be experiencing pure shear,

while if it is inclined on some plane respect to the longitudinal axis of the shaft, it will be
experiencing normal stresses (tensile on two faces and compressive on the other two
faces) and shearing stresses.

Now it is time to analyze the four types of shafts described in the beginning of the
chapter. We will start with the uniform shaft. When talking about strains, we have seen
that the equation to calculate the maximum shearing strain in a uniform shaft of length L
and radius c is
L
c
=
max
, where is the angle of twist. Another thing we have seen is
that the equation that relates the torque with the maximum shearing stress is
J
c T
=
max
.
If despite of the torque we are applying to the shaft it still continues working in the elastic
range (which means that the maximum shearing stress does not exceed the yield stress),
then Hookes law can be applied, and we would get that
G
max
max

= . From here,
through a few substitutions
4
with the equations presented before, we can get
that
J G
c T

=
max
, and if we solve for the angle of twist, we get that

4
Beer & Johnston et. al., Mechanics of Materials, McGraw-Hill, New York, Third Edition, 2001, p. 150
J G
L T

=
Here we can see the relationship between the angle of twist and the torque, the length of
the shaft, the modulus of rigidity and the polar moment of inertia. In all of these cases,
the angle of twist is proportional to the other factors.

In the case of a multiple-section shaft, we can calculate the angle of twist as stated
before by making cut planes in every section of the shaft, and then applying the principle
of static equilibrium. In this way, we will be able to determine which the angle of twist of
each section is, and the total angle of twist is going to be the summation of all the of the
sections. In each of these sections, the four parameters to calculate the angle of twist are
not going to be constant. If the shaft is made of two materials, then it wont have the
same modulus of rigidity. If different torqueses are applied to each section of the shaft,
then the torque is not going to be constant either. Also, the length can change, as well as
the polar moment of inertia, which varies depending of the radius of the section. So, if we
want to express the total angle of twist of a shaft including n number of sections, then it
will be expressed as

=
n
i i i
i i
J G
L T
1

To study how the angle of twist is in a variable cross section shaft, we must first
take a look at Figure 20. As the cross section is varying along the axis, we must make an
integration of all the small elements contained in the shaft. For this, take a



Figure 20 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

small disc (which in Figure 20 is taken as dx). This small portion of the shaft is going to
be taken as the length of a small disc. Once this is done, the equation for the angle of
twist would be
J G
dx T
d

= . It is important to understand that the polar moment of inertia

is a function of x (at each different point of the shaft the radius is different), and that this
function has to be determined. So, if we want to obtain the total angle of twist of the
shaft, we have to make a summation of each small disk (by making an integral). This
integral has to be over the length of the shaft (which gives us the limits of the integral).
The expression to calculate the total angle of twist is

=
L
J G
dx T
0

An alternative form of this equation is

=
L
x J x G
dx x T
0
) ( ) (
) (
, where we can see that each term
of the equation is a function of x.

Finally, we have one of the most common cases of application of shafts: shafts
with relative rotation. By this concept, we understand that both ends of the shaft are
rotating. In this case the angle of twist of the shaft is the angle through which one end
rotates with respect to the other. (Beer & J ohnston, 2001). Consider Figure 21. By
assuming that each shaft is elastic, and that they have length L, radius c, and modulus of
rigidity G, if a torque

Figure 21 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

T is applied at A, then both shafts will start rotating. As seen in the Figure, the end D is
fixed, so this will cause an angle of twist which will be measured by the rotation
c
of
end C. In the other shaft, both ends are rotating, so the rotation of the shaft is going to be
the difference between the angles of rotation
b
and
a
. This is known as relative rotation.
Knowing this, we can get the equation for relative rotation:
J G
L T
B A B A

= =
/

To end this chapter, Figure 22 is going to be presented, where we can see an
example of how each kind of shaft is: uniform, multi-section, variable cross-section or
relative rotation. Each case has been analyzed, and this Figure is only to understand
which the composition of each case is: which ends are fixed, which ends are rotating,
how is the cross-section of the shaft along the longitudinal axis, or if it is transmitting
torque from one shaft to another with power transmission elements like gears, spurs,
pulleys, etc.

Figure 22 (Beer & J ohnston et. al. Interactive Tutorial, 2001)

Safety Factor
An important thing to consider when designing any engineering element is the
safety factor. The maximum load that a structural member or a machine component will
be allowed to carry under normal conditions of utilization is considerably smaller than
the ultimate load. (Beer & J ohnston, 2001). This smaller load is known as allowable
load, working load or design load. So, when an element is working (transmitting torque,
for example), it is going to be working safely until a limit (the allowable load), and only a
portion of the ultimate-load capacity is going to be used. The remaining part of the
capacity of the element to support load is kept in reserve to assure that the element is
working safely. The ratio of the ultimate load to the allowable load is used to define the
factor of safety. So the equation to calculate it is
Factor of safety =
load allowable
load ultimate
S F

. . =

To choose the correct factor of safety, some considerations are required
5
:
1. Variations that may occur in the properties of the member under consideration.
2. The number of loadings that may be expected during the life of the structure or
machine.
3. The type of loadings that are planned for in the design or that may occur in the
future.
4. The type of failure that may occur.
5. Uncertainty due to methods of analysis.
6. Deterioration that may occur in the future because of poor maintenance or
because of unpreventable natural causes.
7. The importance of a given member to the integrity of the whole structure.

The factor of safety for shafts is 3. This factor was chosen by considering the
variations in material properties, the effects of size of the shafts, the type of loading, the
effect of machining or forming, the effect of heat treatment, the effect of shear on
function, the effect of operating environment, the specific operating requirements and the
concern for the human safety.


5
Taken from Beer & J ohnston et. al., Mechanics of Materials, McGraw-Hill, New York, Third Edition,
2001, p. 29