DESIGN FOR DIFFERENT TYPES

OF LOADING
5–2 TYPES OF LOADING AND STRESS RATIO
The primary factors to consider when specifying the type of loading to which a
machine part is subjected are the manner of variation of the load and the
resulting variation of stress with time.
Stress variations are characterized by four key values, expressed here as normal
stresses:
1. Maximum stress, 𝜎𝑚𝑎𝑥
2. Minimum stress, 𝜎𝑚𝑖𝑛
3. Mean (average) stress, 𝜎𝑚
4. Alternating stress, 𝜎𝑎 (stress amplitude)
The maximum and minimum stresses are usually computed from known
information by stress analysis or finite-element methods, or they are measured
using experimental stress analysis techniques.
Then the mean and alternating stresses can be computed from

The behavior of a material under varying stresses is dependent on the manner of

the variation.
One method used to characterize the variation is called stress ratio.
Two types of stress ratios that are commonly used are defined as follows:

Stress ratio R is used in this book.

Static Stress
When a part is subjected to a load that is applied slowly, without shock, and is held
at a constant value, the resulting stress in the part is called static stress.
An example is the load on a structure due to the dead weight of the building
materials.
Figure 5–1 shows a diagram of stress versus time for static loading.
Because 𝜎𝑚𝑎𝑥 = 𝜎𝑚𝑖𝑛 , the stress ratio for static stress is R = 1.0.
Static loading can also be assumed when a load is applied and is removed slowly
and then reapplied, if the number of load applications is small, that is, under a few
thousand cycles of loading.
Repeated and Reversed Stress—Pure Oscillation
A stress reversal occurs when a given element of a load-carrying member is
subjected to a certain level of tensile stress followed by the same level of
compressive stress.
If this stress cycle is repeated many thousands of times, the stress is called
repeated and reversed or pure oscillation.
Figure 5–2 shows the diagram of stress versus time for repeated and reversed
stress.
Because 𝜎𝑚𝑖𝑛 = −𝜎𝑚𝑎𝑥 , the stress ratio is R = -1.0, and the mean stress is zero.
An important example in machine design is a rotating circular shaft loaded in
bending such as that shown in Figure 5–3.
In the position shown, an element on the bottom of the shaft experiences tensile
stress while an element on the top of the shaft sees a compressive stress of equal
magnitude.
As the shaft is rotated 180° from the given position, these two elements
experience a complete reversal of stress.
Now if the shaft continues to rotate, all parts of the shaft that are in bending see
repeated, reversed stress.
This is a description of the classic loading case of repeated and reversed bending.
This type of loading is often called fatigue loading, and a machine of the type
shown in Figure 5–3 is often used to test the fatigue behavior of materials.
The device shown is called the R. R. Moore fatigue test device and the material
property thus measured is called endurance limit and this property is discussed in
detail in Section 5–5.
The shaft is supported by a bearing at each end while a yoke is supported on
bearings.
A known loading applied to the yoke resulting in two concentrated loads being
applied; one at each bearing that supports the yoke.
Note from the shearing force and bending moment diagrams that this type of
loading provides uniform bending moment between the yoke arms while the
shearing force is zero.
Thus pure bending occurs in the test section.
The shaft is machined to precise dimensions with the middle portion having a very
gradual taper down to a small diameter.
That diameter is typically 0.300 in.
With the gradual taper, the stress concentration factor is virtually 1.0.
Furthermore, the shaft is polished to a fine surface finish so that machining marks
do not affect the stress levels in the bar.
The shaft is rotated by an electric motor while the system counts the number of
revolutions.
It also has a device to sense when the specimen breaks so that there is a known
relationship between the stress level and the number of cycles to failure.
Actually, reversed bending is only a special case of fatigue loading, since any stress
that varies with time can lead to fatigue failure of a part.
Many materials test laboratories are using computer-controlled, repeated and
reversed axial loading instead of rotating bending to acquire fatigue strength data.
It is described later that there are differences between these two methods in
regard to the strength values obtained.
It is essential that care be exercised to determine what type of stress is used to
measure the fatigue strength when using published data.
Fluctuating Stress—Pulsating Stress
When a load-carrying member is subjected to an alternating stress with a nonzero
mean, the loading produces fluctuating stress, sometimes called pulsating
stress. Figure 5–4 shows four diagrams of stress versus time for this type of stress.
Differences among the four diagrams occur in whether the various stress levels are
positive (tensile) or negative (compressive).
Any varying stress with a nonzero mean is considered a fluctuating stress.
Figure 5–4 also shows the possible ranges of values for the stress ratio R for the
given loading patterns.
A special, frequently encountered case of fluctuating stress is repeated, one-
direction stress, or pure pulsating stress, in which the load is applied and removed
many times.
As shown in Figure 5–5, the stress varies from zero to a maximum with each cycle.
Then, by observation,

An example of a machine part subjected to the more general nature of fluctuating

stress of the type shown in Figure 5–4(a) is shown in Figure 5–6, in which a
reciprocating cam follower feeds spherical balls one at a time from a chute.
The follower is held against the rotating eccentric cam by a flat spring loaded as a
cantilever.
Part (a) of the figure shows the entire layout of the ball feed device and part (b)
shows the cross section of the flat spring.
Parts (c) and (d) show two views of just the cam,follower, and flat spring.
When the follower is farthest to the left, the spring is deflected from its free
(straight) position by an amount ymin = 3.0 mm.
When the follower is farthest to the right, the spring is deflected to
ymax = 8.0 mm.
Then, as the cam continues to rotate, the spring sees the cyclic loading between
the minimum and maximum values.
Point A at the base of the spring on the convex side experiences the varying tensile
stresses of the type shown in Figure 5–4(a).
Example Problem 5–1 completes the analysis of the stress in the spring at point A.
Example Problem 5–1
For the flat steel spring shown in Figure 5–6, compute the maximum stress, the
minimum stress, the mean stress, and the alternating stress. Also compute the
stress ratio, R. The length L is 65 mm. The dimensions of the spring cross section
are t = 0.80 mm and b = 6.0 mm.

Solution
Objective Compute the maximum, minimum, mean, and alternating tensile
stresses in the flat spring. Compute the stress ratio, R.
Given Layout shown in Figure 5–6. The spring is steel: L = 65 mm.
Spring cross-sectional dimensions: t = 0.80 mm and b = 6.0 mm.
Maximum deflection of the spring at the follower = 8.0 mm.
Minimum deflection of the spring at the follower = 3.0 mm.
Analysis Point A at the base of the spring experiences the maximum tensile
stress. Determine the force exerted on the spring by the follower for
each level of deflection using the formulas from Table A14–2, Case (a).
Compute the bending moment at the base of the spring for each
deflection.
Then compute the stresses at point A using the bending stress
formula, s = Mc/I.
Use Equations (5–1), to (5–3) for computing the mean, alternating
stresses, and R.
Results Case (a) of Table A14–2 gives the following formula for the amount of
deflection of a cantilever for a given applied force:

Solve for the force as a function of deflection:

Appendix 3 gives the modulus of elasticity for steel to be E = 207 GPa.
The moment of inertia, I, for the spring cross section is found from

Then the force on the spring when the deflection y is 3.0 mm is

The bending moment at the support is

The bending stress at point A caused by this moment is

This is the lowest stress that the spring sees in service, and therefore
Because the force on the spring is proportional to the deflection, the
force exerted when the deflection is 8.0 mm is

The bending moment is

The bending stress at point A is

This is the maximum stress that the spring sees, and therefor

Now the mean stress can be computed:

Finally, the alternating stress is

The stress ratio is found using Equation (5–3):

The sketch of stress versus time shown in Figure 5–4(a) illustrates the
form of the fluctuating stress on the spring.
In Section 5–8, you will see how to design parts subjected to this kind
of stress
Shock or Impact Loading
Loads applied suddenly and rapidly cause shock or impact.
Examples include a hammer blow, a weight falling onto a structure, and the action
inside a rock crusher.
The design of machine members to withstand shock or impact involves an analysis
of their energy-absorption capability, a topic not considered in this book. (See
References 3, 6, 10, and 12.)
Random Loading
When varying loads are applied that are not regular in their amplitude, the loading
is called random.
Statistical analysis is used to characterize random loading for purposes of design
and analysis.
This topic is not covered in this book. See Reference 11.
5–3 FAILURE THEORIES
When a part is designed, it is of primary interest to know if the part will fail during
service.
As described in Section 4–3, one of the design objectives is to ensure that the
material has the strength to sustain the stress.
Since materials behave differently under different types of loading conditions,
various failure theories have been proposed and tested.
The theories presented here are accepted practices for designers to compare some
defined stresses to some defined strengths.
For ductile materials under static loading, a stress element is considered to have
failed when yielding occurs.
Thus, the failure theories are based on established yield criteria.
As brittle materials do not exhibit yielding before fracture, the evaluation of failure
is based on postulated fracture criteria.
For cyclic loading, other failure theories for evaluating the mean and alternating
stresses have been proposed.
In Section 5–4, we will introduce the failure theories for static loading. Sections 5–5
and 5–6 present the concepts of fatigue strength and endurance limit that are
critical in designing parts for cyclic loading.
The theories for fatigue failure are presented in Section 5–7.
5–4 DESIGN FOR STATIC LOADING
Ductile Materials under Static Loading
There are two widely accepted ductile material failure theories for parts under
static loading: Maximum Shear Stress Theory and Distortion Energy Theory. These
theories predict failure due to yielding that produce permanent, plastic
deformation.
Maximum Shear Stress Theory (MSST).
The MSST of failure prediction states that a ductile material begins to yield when
the maximum shear stress at a critical location in a load-carrying component
exceeds that in a tensile-test specimen when yielding begins.
A Mohr’s circle analysis for the uniaxial tension test, discussed in Chapter 4, shows
that the maximum shear stress is one half of the applied tensile stress.
At yield, then, 𝑠𝑠𝑦 = 𝑠𝑦 Τ2.
We use this approach in this book to estimate 𝑠𝑠𝑦 .
Then, for design, the stress element at the location of interest is safe when:

where N is the design factor (to be discussed in Section 5–9).

As shown in Section 4–4, the maximum shear stress is the Tresca stress:

Thus, because 𝑠𝑠𝑦 = 𝑠𝑦 Τ2 ,

where 𝜎1 and 𝜎3 are the maximum and minimum principal stresses.

Both Equations (5–5) and (5–6) evaluate the diameter of the Mohr’s circle against
the yield strength of the material.
When considering all three principal stresses, two additional cases should also be
noted.
For the condition that all of the principal stresses are in tension, 𝜎1 > 𝜎2 > 𝜎3 , the
stress element is safe when

For the condition that all of the principal stresses are in compression, 0 > 𝜎1 > 𝜎2
> 𝜎3 , the stress element is safe when

This concept can be illustrated by the yield locus shown in Figure 5–7.
The numerical scales on the graph are normalized to the yield strength,
𝜎1 Τ𝑠𝑦 = 1.0 .
The materials are even materials for which the values of yield strength in tension
and compression are the same, a characteristic of ductile materials.
Plotting the principal stresses, the stress states that lie within the hexagon are
predicted to be safe, while those outside would predict failure.
The maximum shear stress method of failure prediction has been shown by
experimentation to be somewhat conservative.
It is relatively easy to use and is often chosen by designers.
For more precise analysis, the distortion energy method is preferred.
Distortion Energy Theory (DET).
The distortion energy Theory has been shown to be the best predictor of failure
for ductile materials under static loads.
It is applied using the von Mises stress, 𝜎𝑒 , which was described in Section 4–4.
The von Mises stress can be presented in different forms, including the general
form for 3D stress element:
A stress element is considered safe when

A number of variations based on 2D stress state can be derived from Equation (5–
9).
For example, the von Mises stress of the stress element 𝜎1 > 𝜎2 > 0 and 𝜎3 > 0 is
expressed as

It is also helpful to visualize the distortion energy failure prediction theory by

plotting the von Mises yield locus on a graph, as shown in Figure 5–7.
The yield locus is an ellipse centered at the origin and passing through the yield
strength on each axis, in both the tensile and compressive regions.
As the concept demonstrated in the MSST, the stress states that lie within the
ellipse are predicted to be safe, while those outside would predict failure.
The comparison of the MSST and DET is shown in Figure 5–7.
With data showing that the distortion energy theory is the best predictor, it can be
seen that the MSST is generally conservative and that it coincides with the
distortion energy ellipse at six points.
In other regions, it is as much as 16% lower.
Note the 45° diagonal line through the second and fourthquadrants, called the
pure shear diagonal.
It was demonstrated in Section 4–6 that for pure shear the stress state is

Substituting into Equation (5–9):

and thus

for N = 1. This predicts yielding when the shear stress is 0.577𝑠𝑦 .

The MSST predicts failure at 0.50𝑠𝑦 , thus quantifying the conservatism of the
MSST.
Brittle Materials under Static Loading
Brittle materials do not yield. Thus, their failure prediction is based on fracture
criteria.
Brittle materials are typically “non-even” meaning that their compressive strength
is higher than the tensile strength.
This characteristic is reflected in the three failure theories that are presented here.
Maximum Normal Stress Theory (MNST).
The MNST states that a material will fracture when the maximum normal stress
(either tension or compression) exceeds the ultimate strength of the material as
obtained from a standard tensile or compressive test.
Its use is limited, namely for brittle materials under pure uniaxial static tension or
compression.
When applying this theory, any stress concentration factor,𝑲𝒕 , at the region of
interest should be applied to the computed stress because brittle materials do not
yield and therefore cannot redistribute the increased stress.
Given the principal stresses, 𝜎1 > 𝜎2 > 𝜎3 , the following equations apply the
maximum normal stress theory for safe design.

where 𝑠𝑢𝑡 and 𝑠𝑢𝑐 are the uniaxial ultimate strength in tension and compression,
respectively.
It should be noted that the value of 𝑠𝑢𝑐 is negative.
Figure 5–8 shows the failure line defined by MNST.
The stress states within the square area are predicted to be safe.
Coulomb-Mohr Theory (CMT).
As discussed in the previous section, the MNST is useful for failure prediction of
brittle materials subjected to uniaxial tension and uniaxial compression.
For a biaxial stress element, a failure criterion known as the Coulomb-Mohr
Theory can better predict failure by fracture.
The following equation expresses the basis for the CMT:

The fracture prediction based on the CMT is an uneven hexagon as shown in

Figure 5–8.
Equation (5–16) represents the line connecting the points (𝜎𝑢𝑡 , 0) and (0, 𝜎𝑢𝑐 ).
It can be observed that the CMT is an extension of MSST with consideration of the
non-even material property of brittle materials. Similar to MSST, when
considering all three principal stresses, two additional cases should also be
evaluated.
When all of the principal stresses are in tension, 𝜎1 > 𝜎2 > 𝜎3 , the stress
element is safe when

When all of the principal stresses are in compression, 0 > 𝜎1 > 𝜎2 > 𝜎3 , the stress
element is safe when

Note that the value of 𝜎𝑢𝑐 is negative.

Modified Mohr Theory (MMT).
The Modified Mohr Theory is a semi-empirical data fitting method that best
predicts fracture of brittle material under static loading, especially for the stress
states in the fourth quadrant as depicted in Figure 5–8 where the circles illustrate
examples of experimental results.
Failure is predicted when the stress state is outside of the region defined by
connecting the points (𝜎𝑢𝑡 , 𝜎𝑢𝑡 ), (−𝜎𝑢𝑡 , 𝜎𝑢𝑡 ), (𝜎𝑢𝑐 , 0 ), (𝜎𝑢𝑐 , 𝜎𝑢𝑐 ), (0, 𝜎𝑢𝑐 ),
(𝜎𝑢𝑡 , −𝜎𝑢𝑡 ), and (𝜎𝑢𝑡 , 𝜎𝑢𝑡 ).
For design, because of the many different shapes and dimensions of safe-stress
zones, it is suggested that a rough plot be made of the pertinent part of the
modified Mohr diagram from actual material strength data as shown in Figure 5–9.
Then the actual values of 𝜎1 and 𝜎2 can be plotted to ensure that they lie within
the safe zone of the diagram as shown in Figure 5–9.
A load line can be an aid in determining the design factor, N, using the modified
Mohr diagram.
The assumption is made that stresses increase proportionally as loads increase.
Apply the following steps for an example stress state, A, for which 𝜎1𝐴 = 15𝑘𝑠𝑖 and
𝜎2𝐴 = −80𝑘𝑠𝑖. The material is Grade 40A gray cast iron having 𝜎𝑢𝑡 = 40𝑘𝑠𝑖
and𝜎𝑢𝑐 = −140𝑘𝑠𝑖.
1. Draw the modified Mohr diagram as shown in Figure 5–9.
2. Plot point A (15, -80 ).
3. Draw the load line from the origin through point A until it intersects the failure
line on the diagram at the point labeled Af.
4. Determine the distances OA = 81.4 ksi and OAf = 96.0 ksi by scaling the diagram.
5. Compute the design factor from N = OAf/OA = 96.0/81.4 = 1.18.
Summary of Static Loading Failure Theories
It is shown in this section that the selection of static loading failure theories mainly
depends on the material (ductile or brittle).
The information needed for failure prediction includes the principal stresses and
the yield strength or ultimate tensile/compressive strength as shown in the various
equations.
A summary of the presented failure theories is shown in Table 5–1.
5–5 Endurance Limit And Mechanisms Of Fatigue Failure
Whenever a machine element is subjected to cyclical loading characterized by the
patterns like those shown in Figures 5–2, 5–4, and 5–5, the loading is generally
called fatigue loading, common to machinery components.
Yield strength and ultimate tensile strength of a material are not adequate to
represent the ability of materials to resist fatigue loading.
This section presents the concept of endurance limit, sometimes called fatigue
limit, that must be used in such cases.
Components may fail at stress levels lower than ultimate or yield strengths after
experiencing applied stresses for several cycles.
Fatigue failures are often classified as either low-cycle fatigue (LCF) or high-cycle
fatigue (HCF) because the mechanism of failure is different for each.
While no specific dividing line can be defined, designers often use up to 1000
cycles (103) for LCF and higher numbers of cycles—up to infinite life—as HCF.
Fatigue failures start at small surface cracks, internal imperfections, or even at
grain boundaries in the material in areas subjected to tensile stress.
With repeated applications of stress, the cracks grow and progress to larger areas
of the cross section.
Eventually, the component fails, often suddenly and catastrophically.
Such failures frequently occur within areas of stress concentration such as keyways
or grooves in shafts, steps in the size of a cross section, notches, or other
geometric discontinuities as discussed in Section 3–22.
Even surface roughness from machining or accidental nicks and scratches can serve
as points of crack initiation.
Therefore, designers must consider the possibility of fatigue failure when sizing
critical sections of components.
Manufacturers must also understand this phenomenon and produce parts with
good finishes that are free from damage.
End users of critical components must also handle them with care.
In low-cycle fatigue, local stresses experience high strain levels, approaching or
exceeding the strain at yield of the material.
Such events may be due to accidental overloading, or infrequently encountered
situations during fabrication of a component, installation into an assembly, shock
during transportation or handling, evasive maneuvers, takeoffs or landings of
aircraft, launch of a ship or spacecraft, initial testing, seismic shock during an
earthquake, or operating for prolonged periods near the limits of the capability of
a system.
The high strain may cause microscopic cracks that progress to ultimate failure.
Prediction of the life of a component under such conditions falls under the analysis
procedure called fracture mechanics that requires extensive knowledge of the
geometry of the crack and the ability to characterize how a specific material
behaves in the highly localized region of high strain and stress around the crack.
A life prediction method called strain-life is used. References 1, 2, 7, and 9 and
Internet sites 1–8 contain extensive detail about these topics.
The endurance limit of a material under high-cycle fatigue loading is determined
from tests that apply cyclic patterns of stress for long periods of time, and data are
obtained for the number of cycles to failure for a given stress level.
As expected, higher stress levels produce failure at fewer numbers of cycles and
lower stresses permit higher numbers of cycles—up to a point.
For many common materials used in machinery, a stress level is reached where a
virtually unlimited number of cycles of stress can be applied without fatigue
failure.
This stress level is called the endurance limit or fatigue limit of the material.
In this book, we use the symbol 𝑠𝑛 for this property.
Data for endurance limit are reported on charts such as that shown in Figure 5–10,
called a stress-life diagram.
The vertical axis is the stress amplitude, 𝜎𝑎 ,as defined by Equation (5–2) and
shown in Figure 5–2, and it is assumed that the same amplitude occurs for each
cycle for many thousands of cycles.
The horizontal axis is the number of cycles to failure, N.
Both axes are logarithmic scales, resulting in the data plotting as straight lines.
The data are average values of endurance limit through the scattered data points
taken from numerous tests at each stress level.
The transition from the sloped line to the horizontal line at the fatigue limit for any
given material typically occurs at approximately one million cycles (106), and the
curves shown are idealized, showing the break to be sharp.
The following equation represents the sloped portion of the curve:
Data for these properties for many materials are presented in References 1, 2, 7,
and 9, and Internet sites 1–8 describe software programs that contain sizable
databases of such data.
Table 5–2 shows the data for the five selected materials shown in Figure 5–10 for
one plain carbon steel, two alloy steels, and two aluminum alloys, taken from
Internet site 1.
The last column for curve intercept, 𝑠𝑓ƴ , represents the stress value where the curve
intersects the vertical axis.
This value has no further use as discussed later for low-cycle fatigue.
Note the difference between the curves for the three steels and those for the two
aluminum alloys.
The steels exhibit a true fatigue limit resulting in the horizontal line to the right of
106 cycles, and should never fail by fatigue at higher numbers of cycles of loading.
The curves for aluminum continue to drop after 106 cycles, although at a much
reduced slope.
Therefore, data for endurance limits of aluminum, many other nonferrous metals,
and some very-high-strength ferrous metals are quoted as a value of 𝑠𝑛 at a stated
number of cycles, typically 106 or 107.
For higher numbers of cycles, additional data should be sought.
The rotating bending test, as shown in Figure 5–3, has been used for many years to
acquire endurance limit data and much of the reported data are based on this test.
The specimen has a small diameter (typically 0.30 in or 7.62 mm) and is highly
polished to eliminate any effect of surface texture.
The shape and manner of loading produces pure bending with zero shearing stress
and no stress concentrations in the central section.
The magnitude of the load can be varied to produce a desired stress level and the
shaft is rotated until it breaks.
The total number of revolutions to failure is recorded.
The test produces the classic repeated and reversed stress shown in Figure 5–2
having a mean stress of zero, stress amplitude of 𝑠𝑎 , and stress ratio, R = -1.0.
In recent times, other methods have gained favor, particularly programmable
servo-controlled axial tension testing devices.
Test specimens can be loaded in many different patterns, simulating any of the
conditions shown in Figures 5–2, 5–4, 5–5, and others.
When the load is reversed and repeated, seemingly similar to the rotating bending
test, the stress cycle shown in Figure 5–2 is produced, resulting in a mean stress of
zero and a stress ratio, R = -1.0.
However, an important difference in the behavior of the material occurs because
the stress distribution created in the specimens is ideally uniform across the entire
section.
Note that for rotating bending, only the outermost part of the cylindrical specimen
sees the maximum stress and the stress decreases linearly to zero at the center of
the bar.
Fatigue failures are more likely to initiate in regions of high tensile stress.
Because in the axial load test all of the material is subjected to the highest stress,
reported endurance limit data are typically lower than those for the rotating
bending test by approximately 20%.
This situation is discussed more in Section 5–6.
The servo-controlled testing devices are also used to evaluate the effect of loading
that produces a fluctuating or pulsating cyclical stress pattern with a nonzero mean
stress such as those shown in Figures 5–4 and 5–5.
The mean stress effect is discussed later.
References 4 and 11 include many tables of data for fatigue strengths of materials
along with additional detail on the nature of fatigue failures.
Data for the endurance limit should be used wherever it is available, either from
test results or from reliable published data.
However, such data are not always available.
Reference 5 suggests the following approximation for the basic rotating bending
endurance limit for wrought steel having 𝑠𝑢 ≤ 1500𝑀𝑃𝑎 220𝑘𝑠𝑖 .
Low-cycle Fatigue
That part of Figure 5–10 to the left of the line for 1000 cycles is the low-cycle
fatigue region and the discussion above does not apply.
In fact, the parts of the lines from N = 1000 down to N = 1 are not used at all,
except for providing a convenient way of drawing the sloped part of the curves—a
straight line from the curve intercept to the fatigue limit.
Failure at a single cycle (N = 1), of course, occurs at the ultimate tensile strength for
the material, 𝑠𝑢 .
Some designers add a supplementary line from N = 1000 to N = 1, as shown
in dashed form for 4340 steel.
However, it is recommended that the strain-life technique mentioned earlier
be used in this region.
5–6 Estimated Actual Endurance Limit, 𝒔𝒏ƴ
If the actual material characteristics or operating conditions for a machine part are
different from those for which the basic endurance limit was determined, the
fatigue strength must be reduced from the reported value.
Some of the factors that decrease the endurance limit are discussed in this section.
The discussion relates only to the endurance limit for materials subjected to
reversed and repeated bending stress.
Cases involving endurance limit in shear are discussed separately in Section 5–11.
We begin by presenting a procedure for estimating the actual endurance limit, 𝑠𝑛ƴ ,
for the material for the part being designed.
It involves applying several factors to the basic endurance limit for the material.
Additional elaboration on the factors follows.
PROCEDURE FOR ESTIMATING ACTUAL ENDURANCE LIMIT, 𝒔𝒏ƴ ▼
1. Specify the material for the part and determine its ultimate tensile strength, 𝑠𝑢 ,
considering its condition, as it will be used in service.
2. Specify the manufacturing process used to produce the part with special
attention to the condition of the surface in the most highly stressed area.
3. Use Figure 5–11 to estimate the modified endurance limit, 𝑠𝑛 .
4. Apply a material factor, 𝐶𝑚 , from the following list.
Wrought steel: 𝐶𝑚 = 1.00
Cast steel: 𝐶𝑚 = 0.80
Powdered steel: 𝐶𝑚 = 0.76
Malleable cast iron: 𝐶𝑚 = 0.80
Gray cast iron: 𝐶𝑚 = 0.70
Ductile cast iron: 𝐶𝑚 = 0.66
5. Apply a type-of-stress factor: 𝐶𝑠𝑡 = 1.0 for bending stress; 𝐶𝑠𝑡 = 0.80 for axial
tension.
6. Apply a reliability factor, CR, from Table 5–3.
7. Apply a size factor, Cs, using Figure 5–12 and Table 5–4
as guides.
8. Compute the estimated actual endurance limit, 𝑠𝑛ƴ ,
from
These are the only factors that will be used consistently in this book. If data for
other factors can be determined from additional research, they should be
multiplied as additional terms in Equation (5–21).
In most cases, we suggest accounting for other factors for which reasonable data
cannot be found by adjusting the value of the design factor as discussed in Section
5–9.
Stress concentrations caused by sudden changes in geometry are, indeed, likely
places for fatigue failures to occur.
Care should be taken in the design and manufacture of cyclically loaded parts to
keep stress concentration factors to a low value.
We will apply stress concentration factors to the computed stress rather than
to the endurance limit. See Section 5–11.
While 12 factors affecting endurance limit are discussed in the following section,
note that the procedure just given includes only the first five. They are surface
finish, material factor, type-of-stress factor, reliability factor, and size factor.
The others are mentioned to alert you to the variety of conditions you should
investigate as you complete a design.
However, generalized data are difficult to acquire for all factors.
Special testing or additional literature searching should be done when conditions
exist for which no data are provided in this book.
The end-of-chapter references contain a huge amount of such information (see
References 2, 4, 7, 9, 11, and 13–16).
Surface Finish
Any deviation from a polished surface reduces endurance limit because the
rougher surface provides sites where locally increased stresses or irregularities in
the material structure promote the initiation of microscopic cracks that can
progress to fatigue failures. Manufacturing processes, corrosion, and careless
handling produce detrimental surface roughening.
Figure 5–11, adapted from data in Reference 8, shows estimates for the endurance
limit 𝑠𝑛 compared with the ultimate tensile strength of wrought steels for
several practical surface conditions.
The data first estimate the endurance limit for the polished specimen to be
0.50 times the ultimate strength and then apply a factor related to the surface
condition. U.S. Customary units are used in Figure 5–11(a) while SI units are shown
in Figure 5–11(b).
Project vertically from the 𝑠𝑢 axis to the appropriate curve and then horizontally to
the endurance limit axis.
The data from Figure 5–11 should not be extrapolated for su 7 220 ksi or 1500
MPa without specific testing as empirical data reported in Reference 4 are
inconsistent at higher strength levels.
Note that the curve labeled Polished is actually the straight line, sn = 0.50su,
implying a factor of 1.0
because endurance limit test specimens are polished.
Ground surfaces are fairly smooth and reduce the endurance limit by a factor of
approximately 0.90 for su 6 160 ksi (1100 MPa), decreasing to about 0.80 for
su = 220 ksi (1500 MPa).
Machining or cold drawing produce a somewhat rougher surface because of
tooling marks resulting in a reduction factor in the range of 0.80 to 0.60 over the
range of strengths shown.
The outer part of a hot-rolled steel has a roughened oxidized scale that produces a
reduction factor from 0.72 to 0.30
if a part is used in the as-rolled condition. For a forged part, not subsequently
machined, the factor ranges from 0.57 to 0.20.
From these data, it should be obvious that you must give special attention to
surface finish for critical surfaces exposed to fatigue loading in order to benefit
from the steel’s basic strength.
Also, critical surfaces of fatigue loaded parts must be protected from nicks,
scratches, and corrosion because they drastically reduce fatigue strength.
Material Factors, Cm
Metal alloys having similar chemical composition can be wrought, cast, or made by
powder metallurgy to produce the final form.
Wrought materials are usually rolled or drawn, and they typically have higher
endurance limit than cast materials.
The grain structure of many cast materials or powder metals and the likelihood of
internal flaws and inclusions tend to reduce their endurance limit.
Reference 5 provides data from which the material factors listed in step 4 of the
procedure outlined previously are taken.
Type-of-Stress Factor, Cst
As discussed in Section 5–6, most endurance limit data are obtained from tests
using a rotating cylindrical bar subjected to repeated and reversed bending in
which the outer part experiences the highest stress. Stress levels decrease linearly
to zero at the center of the bar.
Because fatigue cracks usually initiate in regions of high tensile stress, a relatively
small proportion of the material experiences such stresses.
Contrast this with the case of a bar subjected to direct axial tensile stress for which
all of the material experiences the maximum stress.
There is a higher statistical probability that local flaws anywhere in the bar may
start fatigue cracks.
The result is that the endurance limit of a material subjected to repeated and
reversed axial stress is approximately 80% of that from repeated and reversed
bending.
In this book, we assume that basic endurance limit data are obtained from rotating
bending tests and recommend the factors Cst = 1.0 for bending stress and Cst = 0.80
for axial loading.
Reliability Factor, CR
The data for endurance limit for steel shown in Figure 5–11 represent average
values derived from many tests of specimens having the appropriate ultimate
strength and surface conditions.
Naturally, there is variation among the data points; that is, half are higher and half
are lower than the reported values on the given curve.
The curve, then, represents a reliability of 50%, indicating that half of the parts
would fail.
Obviously, it is advisable to design for a higher reliability, say, 90%, 99%, or
99.9%.
A factor can be used to estimate a lower endurance limit that can be used for
design to produce the higher reliability values. Ideally, a statistical analysis of
actual data for the material to be used in the design should be obtained. By making
certain assumptions about the form of the distribution of strength data, Reference
8 reports the values in Table 5–3 as approximate reliability factors, CR.
Size Factor, Cs—Circular Sections in Rotating Bending
Recall that the basic endurance limit data were taken for a specimen with a circular
cross section that has a diameter of 0.30 in (7.62 mm) and that it was subjected to
repeated and reversed bending while rotating.
Therefore, each part of the surface is subjected to the maximum tensile bending
stress with each revolution.
Furthermore, the most likely place for fatigue failure to initiate is in the zone of
maximum tensile stress within a small distance of the outer surface.
Data from References 5 and 8 show that as the diameter of a rotating circular
bending specimen increases, the endurance limit decreases because the stress
gradient (change in stress as a function of radius) places a greater proportion of
the material in the highly stressed region.
Figure 5–12 and Table 5–4 show the size factor to be used in this book, adapted
from Reference 8. These data can be used for either solid or hollow circular
sections.
Size Factor, Cs—Other Conditions
We need different approaches to determining the size factor when a part with a
circular section is subjected to repeated and reversed bending but it is not
rotating,
or if the part has a noncircular cross section.
Here we show a procedure adapted from Reference 8 that focuses on the volume
of the part that experiences 95% or more of the maximum stress. It is in this
volume that fatigue failure is most likely to be initiated.
Furthermore, in order to relate the physical size of such sections to the size factor
data in Figure 5–12, we develop an equivalent diameter, De. When the parts in
question have a uniform geometry over the length of interest, the volume is the
product of the length and the cross-sectional area.
We can compare different shapes by considering a unit length for each and looking
only at the areas.
As a base, let’s begin by determining an expression for that part of a circular
section subjected to 95% or more of the maximum rotating bending stress, calling
this area, A95.
Because the stress is directly proportional to the radius, we need the area of the
thin ring between the outside surface with the full diameter D and a circle whose
diameter is 0.95D, as shown in Figure 5–13(a). Then,

You should demonstrate that this same equation applies to a hollow circular
section as shown in Figure 5–13(b).
This verifies that the data for size factor shown in Figure 5–12 and Table 5–4 apply
directly to either the solid or hollow circular sections when they experience
rotating bending.
Nonrotating Circular Section in Repeated and Reversed Flexure.
Now consider a solid circular section that does not rotate but that is flexed back
and forth in repeated and reversed bending.
Only the top and bottom segments beyond a radius of 0.475D experience 95% or
higher of the maximum bending stress, as shown in Figure 5–13(c).
By using properties of a segment of a circle, it can be shown that

Now we can determine the equivalent diameter, De, for this area by equating
Equations (5–22) and (5–23) while designating the diameter in Equation (5–21), as
De and then solving for De.

This same equation applies to a hollow circular section. The diameter De can be
used in Figure 5–12 or in Table 5–4 to find the size factor.
Rectangular Section in Repeated and Reversed Flexure.
The A95 area is shown in Figure 5–13(d) as the two strips having a thickness of
0.025h at the top and the bottom of the section.
Therefore

Equating this to A95 for a circular section gives,

This diameter can be used in Figure 5–12 or in Table 5–4 to find the size factor.
Other shapes can be analyzed in a similar manner.
Any Shape in Repeated Direct Axial Tensile Stress.
This special case is based on the concept that the greatest likelihood of initiating a
fatigue failure is in zones of highest tensile stress.
For bending and torsion, the highest tensile stress occurs in the outermost parts of
the cross section and that is the basis for the data in Figure 5–12.
However, for axial tensile stress, all parts of the cross section experience the same
level of tensile stress and, therefore, all parts are equally susceptible to
initiation of a fatigue crack.
For this case, use Cs = 1.0 regardless of the size of the member.
Other Factors
The following factors are not included quantitatively in problem solutions in this
book because of the difficulty of finding generalized data.
However, you should consider each one as you engage in future designs and
seek additional data as appropriate.
Flaws. Internal flaws of the material, especially likely in cast parts, are places in
which fatigue cracks initiate.
Critical parts can be inspected by x-ray techniques for internal flaws.
If they are not inspected, a higher-than average design factor should be specified
for cast parts, and a lower endurance limit should be used.
Temperature. Most materials have a lower endurance limit at high temperatures.
The reported values are typically for room temperatures.
Operation above 500°F (260°C) will reduce the endurance limit of most steels.
See Reference 8.
Nonuniform Material Properties. Many materials have different strength
properties in different directions because of the manner in which the material was
processed.
Rolled sheet or bar products are typically stronger in the direction of rolling than
they are in the transverse direction.
Fatigue tests are likely to have been run on test bars oriented in the stronger
direction. Stressing of such materials in the transverse direction may result in lower
endurance limit.
Nonuniform properties are also likely to exist in the vicinity of welds because of
incomplete weld penetration, slag inclusions, and variations in the geometry of the
part at the weld.
Also, welding of heat-treated materials may alter the strength of the material
because of local annealing near the weld.
Some welding processes may result in the production of residual tensile stresses
that decrease the effective endurance limit of the material.
Annealing or normalizing after welding is often used to relieve these stresses, but
the effect of such treatments on the strength of the base material must be
considered.
Residual Stresses. Fatigue failures typically initiate at locations of relatively high
tensile stress.
Any manufacturing process that tends to produce residual tensile stress will
decrease the endurance limit of the component.
Welding has already been mentioned as a process that may produce residual
tensile stress.
Grinding and machining, especially with high material removal rates, also cause
undesirable residual tensile stresses.
Critical areas of cyclically loaded components should be machined or ground in a
gentle fashion.
Processes that produce residual compressive stresses can prove to be beneficial.
Shot blasting and peening are two such methods.
Shot blasting is performed by directing a high-velocity stream of hardened balls or
pellets at the surface to be treated.
Peening uses a series of hammer blows on the surface. Crankshafts, springs, gears,
and other cyclically loaded machine parts can benefit from these methods.
Corrosion and Environmental Factors.
Endurance limit data are typically measured with the specimen in air.
Operating conditions that expose a component to water, salt solutions, or other
corrosive environments can significantly reduce the effective endurance limit.
Corrosion may cause harmful local surface roughness and may also alter the
internal grain structure and chemistry of the material.
Steels exposed to hydrogen are especially affected adversely.
Nitriding. Nitriding is a surface-hardening process for alloy steels in which the
material is heated to 950°F (514°C) in a nitrogen atmosphere, typically ammonia
gas, followed by slow cooling. Improvement of endurance limit of 50% or more can
be achieved with nitriding.
Effect of Stress Ratio on Endurance Limit.
Figure 5–14 shows the general variation of endurance strength data for a given
material when the stress ratio R varies from -1.0 to +1.0, covering the range of
cases including the following:
■■ Repeated, reversed stress (Figure 5–2); R = -1.0
■■ Partially reversed fluctuating stress with a tensile mean stress [Figure 5–4(b)]; -
1.0 6 R 6 0
■■ Repeated, one-direction tensile stress (Figure 5–5); R = 0
■■ Fluctuating tensile stress [Figure 5–4(a)]; 0 6 R 6 1.0
■■ Static stress (Figure 5–1); R = 1
Note that Figure 5–14 is only an example, and it should not be used to determine
actual data points
If such data are desired for a particular material, specific data for that material
must be found either experimentally or in published literature.
The most damaging kind of stress among those listed is the repeated, reversed
stress with R = -1. (See Reference 4.)
Recall that the rotating shaft in bending as shown in Figure 5–2 is an example of a
load-carrying member subjected to a stress ratio, R = -1.
Fluctuating stresses with a compressive mean stress, as shown in Figures 5–4(c)
and (d), do not significantly affect the endurance limit of the material because
fatigue failures tend to originate in the regions of tensile stress.
Note that the curves of Figure 5–14 show estimates of the endurance limit, sn, as a
function of the ultimate tensile strength for steel.
These data apply to ideal polished specimens and do not include any of the other
factors discussed in this section. For example, the curve for R = -1.0 (reversed
bending) shows that the endurance limit for steel is approximately 0.5 times the
ultimate strength (0.50 * su) for large numbers of cycles of loading (approximately
105 or higher).
This is a good general estimate for steels. The chart also shows that types of
loads producing R greater than -1.0 but less than 1.0 have less of an effect on the
endurance limit.
This illustrates that using data from the reversed bending test is the most
conservative.
We will not use Figure 5–14 directly for problems in this book because our
procedure for estimating the actual endurance limit starts with the use of
Figure 5–11, which presents data from reversed bending tests.
Therefore, the effect of stress ratio is already included. Section 5–7 includes
methods of analysis for loading cases in which the fluctuating stress produces a
stress ratio different from R = -1.0.
Example Problem 5–2
Estimate the actual endurance limit of SAE 1050 cold-drawn steel when used in a
circular shaft subjected to rotating bending only. The shaft will be machined to a
diameter of approximately 1.75 in.
Solution
Objective Compute the estimated actual endurance limit of the shaft material.
SAE 1050 cold-drawn steel, machined.
Size of section: D = 1.75 in.
Type of stress: Reversed, repeated bending.
Given Use the Procedure for Estimating Actual Endurance Limit, 𝑠𝑛ƴ .
Step 1: The ultimate tensile strength: 𝑠𝑢 = 100𝑘𝑠𝑖 from Appendix 3.
Step 2: Diameter is machined.
Step 3: From Figure 5–11, 𝑠𝑛 = 38 ksi.
Step 4: Material factor for wrought steel: Cm = 1.0.
 Step 5: Type-of-stress factor for reversed bending: Cst = 1.0.
Step 6: Specify a desired reliability of 0.99. Then CR = 0.81 (Design
decision).
Step 7: Size factor for circular section with D = 1.75 in.
From Figure 5–12, Cs = 0.83.
Step 8: Use Equation (5–21) to compute the estimated actual
endurance limit.

Comments This is the level of stress that would be expected to produce fatigue
failure in a rotating shaft due to the action of reversed bending. It accounts for the
basic endurance limit of the wrought SAE 1050 colddrawn material, the effect of
the machined surface, the size of the section, and the desired reliability.
Example Problem 5–3
Estimate the actual endurance limit of cast steel having an ultimate strength of 120
ksi when used in a bar subjected to a reversed, repeated, bending load. The bar
will be machined to a rectangular cross section, 1.50 in wide * 2.00 in high.
Solution
Objective Compute the estimated actual endurance limit of the bar material.
Given Cast steel, machined: su = 120 ksi.
Size of section: b = 1.50 in., h = 2.00 in rectangular
Type of stress: Repeated, reversed bending.
Analysis Use the Procedure for Estimating Actual Endurance Limit, 𝑠𝑛ƴ .
Step 1: The ultimate tensile strength is given to be su = 120 ksi.
Step 2: Surfaces are machined.
Step 3: From Figure 5–11, sn = 44 ksi.
Step 4: Material factor for cast steel: Cm = 0.80.
Step 5: Type-of-stress factor for bending: Cst = 1.00.
Step 6: Specify a desired reliability of 0.99. Then CR = 0.81 (Design
decision).
Step 7: Size factor for rectangular section: First use Equation (5–25) to
determine the equivalent diameter,

Then from Figure 5–12, Cs = 0.85.

Step 8: Use Equation (5–21), to compute the estimated actual
endurance limit.
5–7 DESIGN FOR CYCLIC LOADING
Ductile Materials under Cyclic Loading
Similar to the static loading failure theories, prediction of fatigue failure involves
comparison of the stresses to the strengths.
The stresses required are the mean stress and alternating stress introduced in
Section 5–2.
The strengths considered are endurance limit and either yield strength or ultimate
strength of the material.
The concept of fatigue failure theories can be illustrated by introducing the Yield
Line in the 𝜎𝑚 − 𝜎𝑎 coordinate system as shown in Figure 5–15 where the abscissa
is 𝜎𝑚 and the ordinate is 𝜎𝑎 a.
The yield strength of the material is identified on both axes.
By connecting these two points, the Yield Line is defined and it is represented by
the following equation:
A failure theory can be postulated that, when plotting the mean and alternating
stresses of a stress element, the element is safe if the stress state is below the
Yield Line in the triangle region.
Examining the stress state at (𝜎𝑦 , 0) where the mean stress is at the yield strength
and the alternating stress is zero (static loading), the postulated theory seems valid
as in static loading the element is considered safe before the stress reaches the
yield strength.
However, for the stress state at (0, 𝜎𝑦 ), the element is subjected to repeated and
reversed cyclic loading with alternating stress reaching the yield strength.
As it is clear that failure would have occurred when the alternating stress exceeds
the endurance limit of the material, modification of yield line is needed to predict
fatigue failure.
Three empirical fatigue failure criteria are presented below and they are shown in
Figure 5–15.
Soderberg Criterion. The Soderberg criterion gives
a conservative fatigue failure prediction. As shown in Figure 5–15, the Soderberg
line connects 𝜎𝑦 on the 𝜎𝑚 axis to 𝜎𝑛ƴ on the 𝜎𝑎 axis. Its equation is

where Kt is the stress concentration factor and N is the factor of safety. For N = 1,
the safe zone is the triangular area below the Soderberg line.
It is recommended that any stress concentration factor be applied to the
alternating stress component. However, as it is not necessary to apply a stress
concentration factor to the mean stress because the material is ductile and
experimental evidence shows that the presence of a stress concentration does not
affect the contribution of mean stress to fatigue failure.
Goodman Criterion.
The Goodman criterion is perhaps the most-often used analysis for machine
components subjected to cyclic loads. The criterion can be expressed by
connecting the ultimate strength 𝜎𝑢 on the 𝜎𝑚 axis to 𝜎𝑛ƴ on the 𝜎𝑎 axis. Its
equation is

For failure prediction, combinations of mean and alternating stresses that lie under
the Goodman line are considered to be safe.
It can be observed that the main difference between the Goodman and the
Soderberg criterion is the stress states near the lower right-hand side of
the safe zone where the alternating stress is low and the mean stress is beyond the
yield strength of the material.
Therefore, the possibility of yielding must be considered as described next.
Checking for Early Cycle Yielding.
The Goodman line appears to allow a pure mean stress above the yield
strength of the material. While it is possible for ductile material to acquire
additional strength to prevent further yielding (due to work hardening), early cycle
yielding is not commonly acceptable.
As such, an evaluation for early cycle yielding is needed. It is desirable to limit any
possible stress to below the yield strength, and any stress concentration should be
considered for both the alternating and the mean stress to ensure that even
infrequently applied high stresses will not cause damaging strains that could
precipitate fatigue cracks.
Applying a design factor to the yield line results in a design equation to protect
against yield as,
Gerber Criterion.
The Gerber Criterion is a parabolic function that starts at 𝑠𝑢 on the 𝜎𝑚 axis and
ends at 𝑠𝑛ƴ on the 𝜎𝑎 axis:
As can be observed from Figure 5–15, the Gerber line passes generally through the
array of experimental data points (examples of experimental results shown as
circles in the figure) for failures under specific combinations of mean and
alternating stresses.
The starting and end points of the Gerber criterion is the same as that of the
Goodman criterion, and it can be seen that the Goodman line is slightly
conservative.
While the Gerber criterion provides more accurate failure predictions, the
parabolic function is somewhat complicated to implement. See References 5 and 8
for more details on the Gerber Criterion.
Smith Diagram for Showing the Effect of Mean Stress on Fatigue
Some stress analysts, for example many in Germany, use the Smith Diagram to map
a region of acceptable combinations of mean and alternating stresses.
Additional materials property data are required to apply this method as compared
with the Goodman approach defined earlier and used in this book. Figure 5–16
shows one approach to drawing the Smith Diagram for bending stress.
Data required are as follows:
1. Fatigue limit strength for completely reversed and repeated bending stress (𝑅
= −1), called 𝑠𝑛 in this book.
Some call this loading pattern pure oscillation, sn(-1), as shown in Figure 5–2. The
mean stress is zero [𝜎𝑚 = 0] and the stress oscillates between +𝜎𝑚𝑎𝑥 and −𝜎𝑚𝑖𝑛
having the same absolute value.
The alternating stress, 𝜎𝑎 = 𝜎𝑚𝑎𝑥 .
2. Fatigue limit strength for repeated, one-direction bending stress (R = 0),
sometimes called pure pulsating stress, 𝑠𝑛(0) , and shown in Figure 5–5. The
stress oscillates between the value of 𝜎𝑚𝑎𝑥 and 𝜎𝑚𝑖𝑛 = 0.
The alternating stress and the mean stress are equal: 𝜎𝑎 = 𝜎𝑚 = 𝜎𝑚𝑎𝑥 Τ2.
3. Ultimate strength, 𝑠𝑢 .
The steps used to draw the diagram in Figure 5–16 are as follows:
1. Draw the vertical axis for alternating stress, 𝜎𝑎 , and the horizontal axis for mean
stress, 𝜎𝑚 .
2. Draw a dashed line at 45° from the origin. This represents the line of increasing
mean stress.
3. On the vertical axis, plot two points: Point I at +𝜎𝑚𝑎𝑥 = 𝑠𝑛 −1 and Point II at -
− 𝜎𝑚𝑖𝑛 = −𝑠𝑛 −1 for the pure oscillation case.
4. Plot data for the pure pulsating case: Point III at 𝑠𝑛(0) /2,0 and Point IV at 𝑠𝑛(0)
/2, 𝑠𝑛(0) .
5. Draw a line from Point I through Point IV toward its intersection with the 45°
line.
6. Draw another line from Point II through Point III
toward its intersection with the 45° line.
7. Draw a horizontal line from the value of the ultimate strength on the vertical axis
to its intersection with the 45° line, calling that intersection Point V.
8. Call the intersection of the line from Step 7 and the line from Step 5 Point VI.
9. Drop a vertical line from Point VI to its intersection with the line from Step 6,
calling that intersection Point VII.
10. Draw a line from Point V to Point VII.
11. The resulting polygon from Points I → VI → V → VII → II → I encloses an area of
theoretically safe combinations of mean and alternating stresses.
12. The upper part of the polygon from Points I → VI → V is a plot of the maximum
stress that can be applied to a component.
13. The lower part of the polygon from Points V → VII → II is a plot of the minimum
stress corresponding to any maximum stress directly vertically above.
13. The lower part of the polygon from Points
V S VII S II is a plot of the minimum stress
14. For any vertical line drawn through the upper and lower parts of the polygon:
a. The top intersection is +𝜎𝑚𝑎𝑥 .
b. The intersection with the dashed line is 𝜎𝑚 .
c. The lower intersection is 𝜎𝑚𝑖𝑛 .
15. Actual allowable applied stresses, of course, would be lower than these points
according to the desired design factor, N.
The example data used for drawing Figure 5–16 are as
follows:
Ultimate strength = 𝑠𝑢 = 370𝑀𝑃𝑎
Fatigue strength for pure oscillation,𝑠𝑛 −1 = 140𝑀𝑃𝑎
Fatigue strength for pure pulsation,𝑠𝑛(0) = 230𝑀𝑃𝑎 𝑎𝑛𝑑 𝑠𝑛(0) Τ2 = 115𝑀𝑃𝑎
This method is an approximation of data from fatigue tests from many
combinations of mean and alternating stresses.
A modification of this method is to limit any applied stress to the yield strength of
the material instead of the ultimate strength.
Other charts of similar nature are often drawn for torsional shear stresses and
for alternating direct tension–compression.
Brittle Materials under Cyclic Loading
In the past, other than cast iron in compression, brittle materials were not
commonly used for cyclic loading.
Recent work has increased the acceptance of composite and ceramic materials for
fatigue applications.
For brittle materials, stress is significantly increased at locations of stress
concentration and crack tips.
It is recommended to apply a stress concentration factor to both mean and
alternating stresses, similar to the approach of checking early cycle yielding,
Equation (5–29), when using the fatigue failure theory.
Summary of Cyclic Loading Failure Theories
In the three cyclic loading failure theories introduced, the mean and alternating
stresses are presented as 𝜎𝑚 and 𝜎𝑎 , indicating one-dimensional loading.
The failure criteria can be extended to multi-dimensional loading by replacing 𝜎𝑚
and 𝜎𝑎 with 𝜎𝑚ƴ and 𝜎𝑎ƴ :
where 𝜎𝑚ƴ is the mean stress and 𝜎𝑎ƴ is the alternating stress based on Tresca (MSST)
or von Mises (DET) criterion.
The calculation of 𝜎𝑚ƴ and 𝜎𝑎ƴ is accomplished by drawing two Mohr’s circles, one
for the mean stresses and one for the alternating stresses.
From these circles, determine the three principal stresses. Then compute the
“effective stresses” based on MSST or DET for both the mean and the alternating
components.
Since one-dimensional loading is a special case of three-dimensional
loading, Equations (5–27) to (5–30) can be replaced by Equations (5–31) to (5–34).
A summary of the presented fatigue failure theories is given in Table 5–5.
5–8 RECOMMENDED DESIGN AND PROCESSING FOR FATIGUE LOADING
Throughout this chapter, we have described factors that influence the fatigue life
of components subjected to cyclical loads.
This section presents some summary recommendations for designers when
specifying size, shape, and processing techniques for a given component. See
Reference 15 for additional methods and supporting data.
1. Where critical stress levels are encountered, prepare the surface with low
roughness and specify a processing method that will produce well rounded
valleys between peaks. Examples are to use a turning tool for cylindrical parts with
a broad tip radius or a ball end mill with a large diameter for milling operations.
2. Design parts having inherent abrupt changes in geometry with low values of
stress concentrations.
An important example is to provide large radii at sites of change in diameter or
width of a component.
Consult Section 3–22 and Appendix 18.
3. Design the part so that the predominant direction of processing is parallel to the
major axis of the part.
4. Increase the fatigue strength of critical areas of a component by using
processing methods that leave compressive residual stresses. Examples include
shot peening, cold working by rolling or burnishing.
5. Avoid processing methods that produce tensile residual stresses that decrease
the fatigue strength of the material. Examples are (a) aggressive grinding
operations that rapidly remove large amounts of material, (b) some heat-treating
operations that use rapid quenching from high temperatures, and (c) some welding
processes near the heat-affected zone.
If such processes are used, supplementary processing is recommended to relieve
the tensile residual stress.
6. If processing leaves surface irregularities such as burning, cracks, decarburized
materials, or rough welds, take additional steps to remove those irregularities by
light finishing cuts, electropolishing, or lapping.
7. Control the entire path of manufacturing processes to ensure that (a) design
specifications are met; (b) material properties are within the values used in
design analyses; and (c) parts are handled carefully to prohibit accidental surface
damage from nicks, scratches, or corrosion.
5–9 DESIGN FACTORS
The term design factor, N, is a measure of the relative safety of a load-carrying
component.
In most cases, the strength of the material from which the component is to be
made is divided by the design factor to determine a design stress, 𝜎𝑑 , sometimes
called the allowable stress.
Then the actual stress to which the component is subjected should be less than
the design stress.
For some kinds of loading, it is more convenient to set up a relationship from
which the design factor, N, can be computed from the actual applied stresses and
the strength of the material.
Still in other cases, particularly for the case of the buckling of columns, as
discussed in Chapter 6, the design factor is applied to the load on the column
rather than the strength of the material.
Sections 5–4 and 5–7 present methods for computing the design stress or design
factor for several different kinds of loading and materials.
The designer must determine what a reasonable value for the design factor should
be in any given situation.
Often the value of the design factor or the design stress is governed by codes
established by standards setting organizations such as the American Society of
Mechanical Engineers, the American Gear Manufacturers Association, the U.S.
Department of Defense, the Aluminum Association, or the American Institute of
Steel Construction. For structures, local or state building codes often prescribe
design factors or design stresses.
Some companies have adopted their own policies specifying design factors based
on past experience with similar conditions.
In the absence of codes or standards, the designer must use judgment to specify
the desired design factor.
Part of the design philosophy, discussed in Section 5–10, identifies issues such as
the nature of the application, environment, nature of the loads on the component
to be designed, stress analysis, material properties, and the degree of confidence
in data used in the design processes.
All of these considerations affect the decision about what value for the design
factor is appropriate. This book will use the following guidelines.
Ductile Materials
1. N = 1.25 to 2.0. Design of structures under static loads for which there is a high
level of confidence in all design data.
2. N = 2.0 to 2.5. Design of machine elements under dynamic loading with average
confidence in all design data. (Typically used in problem solutions in this book.)
3. N = 2.5 to 4.0. Design of static structures or machine elements under dynamic
loading with uncertainty about loads, material properties, stress analysis, or the
environment.
4. N = 4.0 or higher. Design of static structures or machine elements under
dynamic loading with uncertainty about some combination of loads, material
properties, stress analysis, or the environment.
The desire to provide extra safety to critical components may also justify these
values.
Brittle Materials
5. N = 3.0 to 4.0. Design of structures under static loads for which there is a high
level of confidence in all design data.
6. N = 4.0 to 8.0. Design of static structures or machine elements under dynamic
loading with uncertainty about loads, material properties, stress analysis, or the
environment.
Sections 5–11 and 5–12 provide guidance on the introduction of the design factor
into the design process with particular attention to the selection of the strength
basis for the design and the computation of the design stress. In general, design
for static loading involves applying the design factor to the yield strength
or ultimate strength of the material.
Dynamic loading requires the application of the design factor to the endurance
limit using the methods described in Section 5–6 to estimate the actual endurance
limit for the conditions under which the component is operating.
5–12 DESIGN EXAMPLES
Example design problems are shown here to give you a feel for the application of
the process outlined in Section 5–11.
It is not practical to illustrate all possible situations, and you must develop the
ability to adapt the design procedure to the specific characteristics of each
problem.
Also note that there are many possible solutions to any given design problem. The
selection of a final solution is the responsibility of you, the designer.
In most design situations, a great deal more information will be available than is
given in the problem statements in this book. But, often, you will have to seek out
that information. We will make certain assumptions in the examples that allow the
design to proceed. In your job, you must ensure that such assumptions are
appropriate. The design examples focus on only one or a few of the components of
the given systems. In real situations, you must ensure that each design decision is
compatible with the totality of the design
Design Example 5–1
A large electrical transformer is to be suspended from a roof truss of a building.
The total weight of the transformer is 32 000 lb. Design the means of support.
Solution
Objective Design the means of supporting the transformer.
Given The total load is 32 000 lb. The transformer will be suspended below a
roof truss inside a building. The load can be considered to be static. It
is assumed that it will be protected from the weather and that
temperatures are not expected to be severely cold or hot in the
vicinity of the transformer.
Basic Design Decisions
Two straight, cylindrical rods will be used to support the transformer,
connecting the top of its casing to the bottom chord of the roof truss.
The ends of the rod will be threaded to allow them to be secured by
nuts or by threading them into tapped holes. This design example will
be concerned only with the rods.
 It is assumed that appropriate attachment points are available to
allow the two rods to share the load equally during service.
However, it is possible that only one rod will carry the entire load at
some point during installation. Therefore, each rod will be designed
to carry the full 32 000 lb.
We will use steel for the rods, and because neither weight nor
physical size is critical in this application, a plain, medium-carbon steel
will be used. We specify SAE 1040 cold-drawn steel. From Appendix 3,
we find that it has a yield strength of 71 ksi and moderately high
ductility as represented by its 12% elongation.
The rods should be protected from corrosion by appropriate coatings.
The objective of the design analysis that follows is to determine the
size of the rod.
Analysis The rods are to be subjected to static direct normal tensile stress.
 Assuming that the threads at the ends of the rods are cut or rolled
into the nominal diameter of the rods, the critical place for stress
analysis is in the threaded portion.
Use the direct tensile stress formula, s = F/A. We will first compute
the design stress and then compute the required cross-sectional area
to maintain the stress in service below that value. Finally, a standard
thread will be specified from the data in Appendix Table A2–2(b) for
American Standard threads.
Results In the basic tensile stress equation, s = F/A. The stress state of the 3D
stress element is

From Section 5–4, Design for Static Loading, Equation (5–7) applies for
MSST, the maximum shear stress theory. Letting 𝜎1 = 𝐹 Τ𝐴 = 𝜎𝑑
From Section 5–9, for the design factor, we can specify N = 3, typical for general
machine design with some uncertainty about installation procedures. Then,

Now we can solve for the required cross-sectional area of each rod.

A standard size thread will now be specified. Appendix Table A2–2(b) lists the
1
tensile stress area for American Standard threads. 𝐴 1 − 6𝑈𝑁𝐶 thread (1 1/2 in
2
diameter rod with 6 threads per in) has a tensile stress area of 1.405 in2 which
should be satisfactory for this application.
Comments `The final design specifies a 1 ½ in diameter rod made from SAE 1040
cold-drawn steel with 1 1/2-6 UNC threads machined on each end to
allow the attachment of the rods to the transformer and to the truss.
Note that for this problem, both the MSST [Equation (5–7)] and the
distortion energy theory DET [Equation (5–10)] yield the same result
because the loading is uniaxial tension.
Design Example 5–2
A part of a conveyor system for a production operation is shown in Figure 5–17.
Design the pin that connects the horizontal bar to the fixture. The empty fixture
weighs 85 lb. A cast iron engine block weighing 225 lb is hung on the fixture to
carry it from one process to another, where it is then removed.
It is expected that the system will experience many thousands of cycles of loading
and unloading of theengine blocks.
Solution
Objective Design the pin for attaching the fixture to the conveyor system.
Given The general arrangement is shown in Figure 5–17. The fixture places a
shearing load that is alternately 85 lb and 310 lb (85 + 225) on the pin
many thousands of times in the expected life of the system.
Basic Design Decisions
It is proposed to make the pin from SAE 1020 cold-drawn steel.
 “Design Properties of Carbon and Alloy Steels” lists
𝜎𝑦 51𝑘𝑠𝑖 𝑎𝑛𝑑 𝜎𝑢 = 61𝑘𝑠𝑖.
The steel is ductile with 15% elongation. This material is inexpensive,
and it is not necessary to achieve a particularly small size for the pin.
The connection of the fixture to the bar is basically a clevis joint with
two tabs at the top of the fixture, one on each side of the bar.
There will be a close fit between the tabs and the bar to minimize
bending action on the pin. Also, the pin will be a fairly close fit in the
holes while still allowing rotation of the fixture relative to the bar.
Analysis The Goodman criterion, expressed in Equation (5–32) in Section 5–7,
applies for completing the design analysis because fluctuating
shearing stresses are experienced by the pin.
Therefore, we will have to determine relationships for the mean and
alternating stresses (𝜏𝑚 𝑎𝑛𝑑 𝜏𝑎 ) in terms of the applied loads and the
cross-sectional area of the bar.
Note that the pin is in double shear, so two cross sections resist
the applied shearing force. In general, t = F/2A.
Now we will use the basic forms of Equations (5–1) and (5–2) to
compute the values for the mean and alternating forces on the pin:
The stresses will be found from

From stress transformation, the three mean principal stresses are as follows:

Similarly, the three alternating principal stresses are as follows:

We can now apply the Goodman criterion [Equation (5–32)]

Therefore

The material strength values needed are su = 61 000 psi and sn′. We must find the
value of sn= using the method from Section 5–6.
We find from Figure 5–11 that sn = 23 ksi for the machined pin having a value of su
= 61 ksi.
It is expected that the pin will be fairly small, so we will use Cs = 1.0. The
material is wrought steel rod, so Cm = 1.0. Let’s use Cst = 1.0 to be conservative
because there is little information about such factors for direct shearing stress.
A high reliability is desired for this application,
so let’s use CR = 0.75 to produce a reliability of 0.999 (see Table 5–3).
Then
Because the pins will be of uniform diameter, Kt = 1.0.
Let’s use N = 4 because mild shock is expected. Therefore,

Results Based on the above analysis, we can solve for the required
area, A = 0.03919 in2
Then the required diameter is D = 0.223 in.5Τ16 − 18
Final Design Decisions and Comments
The computed value for the minimum required diameter for the pin, 0.223 in,
5Τ16 − 18 is quite small. Other considerations such as bearing stress and wear at
the surfaces that contact the tabs of the fixture and the bar indicate that a larger
diameter would be preferred.
Let’s specify D = 0.50 in for the pin at this location.
The pin will be of uniform diameter within the area of the bar and the tabs.
It should extend beyond the tabs, and it could be secured with cotter pins or
retaining rings.
This completes the design of the pin. But the next design example deals with the
horizontal bar for this same system.
There are pins at the conveyor hangers to support the bar.
They would also have to be designed.
However, note that each of these pins carries only half the load of the pin in the
fixture connection.
These pins would experience less relative motion as well, so wear should not be so
severe.
Thus, let’s use pins with D = 3/8 in = 0.375 in at the ends of the horizontal bar.
Design Example 5–3
A part of a conveyor system for a production operation is shown in Figure 5–17.
The complete system will include several hundred hanger assemblies like this one.
Design the horizontal bar that extends between two adjacent conveyor hangers
and that supports a fixture at its midpoint.
The empty fixture weighs 85 lb. A cast iron engine block weighing 225 lb is hung on
the fixture to carry it from one process to another, where it is then removed.
It is expected that the bar will experience several thousand cycles of loading and
unloading of the engine blocks.
Design Example 5–2 considered this same system with the objective of specifying
the diameter of the pins.
The pin at the middle of the horizontal bar where the fixture is hung has been
specified to have a diameter of 0.50 in.
Those at each end where the horizontal bar is connected to the conveyor hangers
are each 0.375 in.
Solution
Objective Design the horizontal bar for the conveyor system.
Given The general arrangement is shown in Figure 5–17. The bar is simply
supported at points 24 in apart.
A vertical load that is alternately 85 lb and 310 lb (85 + 225) is applied
at the middle of the bar through the pin connecting the fixture to the
bar.
The load will cycle between these two values many thousands of
times in the expected life of the bar.
The pin at the middle of the bar has a diameter of 0.50 in, while
the pins at each end are 0.375 in.
Basic Design Decisions
It is proposed to make the bar from steel in the form of a rectangular
bar with the long dimension of its cross section vertical.
Cylindrical holes will be machined on the neutral axis of the bar at the
support points and at its center to receive cylindrical pins that will
attach the bar to the conveyor carriers and to
the fixture. Figure 5–18 shows the basic design for the bar.
The thickness of the bar, t, should be fairly large to provide a good
bearing surface for the pins and to ensure lateral stability of the bar
when subjected to the bending stress.
A relatively thin bar would tend to buckle along its top surface where
the stress is compressive.
As a design decision, we will use a thickness of t = 0.50 in.
The design analysis will determine the required height of the bar, h,
assuming that the primary mode of failure is stress due to bending.
Other possible modes of failure are discussed in the comments at the
end of this example.
 An inexpensive steel is desirable because several hundred bars will be
made. We specify SAE 1020 hot-rolled steel having a yield strength of
𝜎𝑦 = 30𝑘𝑠𝑖 and an ultimate strength of 𝜎𝑢 = 55𝑘𝑠𝑖 (Appendix 3)
Analysis The Goodman criterion applies for completing the design analysis
because fluctuating normal stress due to bending is experienced by
the bar. Equation (5–32) will be used:

In general, the bending stress in the bar will be computed from the flexure
formula:
Our approach will be to first determine the values for both the mean
and the alternating bending moments experienced by the bar at its
middle.
Then the yield and endurance limit values for the steel will be found.
Reference 5 in Chapter 3 indicates that a small hole, with diameter d,
in a plate-beam does not weaken the beam if the ratio d/h is less than
0.50. That is, if d/h 6 0.50, Kt = 1.0.
We will make that assumption and check it later.
Based on the application conditions, let’s use N = 4 as advised in item
4 in Section 5–9 because the actual use pattern for this conveyor
system in a factory environment is somewhat uncertain and shock
loading is likely.
Bending Moments Figure 5–18 shows the shearing force and bending
moment diagrams for the bar when carrying just the fixture and then
both the fixture and the engine block.
The maximum bending moment occurs at the middle of the bar
where the load is applied.
The values are Mmax = 1860 lb.in with the engine block on the
fixture and Mmin = 510 lb.in for the fixture alone.
Now the values for the mean and alternating bending moments are
calculated using modified forms of Equations (5–1) and (5–2):

The stresses will be found from

Note that the stress element at the location of interest is subjected

to uniaxial tension/compression without any shear. For 1D loading.
 As such, we can now apply the Goodman criterion from Equation (5–
32) with

Material Strength Values The material strength properties required are the
ultimate strength 𝑠𝑢 and the estimated actual endurance limit
𝑠𝑛ƴ .
 We know that the ultimate strength 𝑠𝑢 = 55𝑘𝑠𝑖. We now find
𝑠𝑛ƴ using the method outlined in Section 5–6.
Size factor, Cs: From Section 5–6, Equation (5–10) defines an equivalent
diameter, De, for the rectangular section as

We have specified the thickness of the bar to be t = 0.50 in.

The height is unknown at this time. As an estimate, let’s assume
h ≈ 2.0 in. Then,

We can now use Figure 5–12 or the equations in Table 5–4

to find Cs = 0.90.
This value should be checked later after a specific height dimension is
proposed.
Material factor, Cm: Use Cm = 1.0 for the wrought, hot-rolled steel.
Stress-type factor, Cst: Use Cst = 1.0 for repeated bending stress.
Reliability factor, CR: A high reliability is desired. Let’s use CR = 0.75 to achieve a
reliability of 0.999 as indicated in Table 5–3.
The value of 𝑠𝑛 = 20𝑘𝑠𝑖 is found from Figure 5–11 for hot-rolled
steel having an ultimate strength of 55 ksi.
Now, applying Equation (5–21) from Section 5–6, we have

Solution for the Requid Section Modulus At this point, we have specified all
factors in Equation (5–32) except the section modulus of the cross
section of the bar that is involved in each expression for stress
as shown above.
` We will now solve the equation for the required value of S.
Recall that we showed earlier that 𝜎𝑚 = 𝑀𝑚 Τ𝑆 𝑎𝑛𝑑 𝜎𝑎 = 𝑀𝑎 Τ𝑆
Then
Results The required section modulus has been found to be S = 0.286 in3.
We observed earlier that 𝑡ℎ2 Τ6 for a solid rectangular cross section,
and we decided to use this form to find an initial estimate for the
required height of the section, h.
We have specified t = 0.50 in. Then the estimated minimum
acceptable value for the height h is

The table of preferred basic sizes in the decimal-inch system (Table

A2–1) recommends h = 2.00 in.
We should first check the earlier assumption that the ratio
𝑑 Τℎ < 0.50 at the middle of the bar.
The actual ratio is
d/h = (0.50 in)/(2.00 in) = 0.25 (okay)
This indicates that our earlier assumption that Kt = 1.0 is correct.
Also, our assumed value of Cs = 0.90 is correct because the actual
height, h = 2.0 in, is identical to our assumed value.
We will now compute the actual value for the section modulus of the
cross section with the hole in it. See Figure A15–6 in the appendix.

This value is larger than the minimum required value of 0.286 in3.
Therefore, the size of the cross section is satisfactory with regard to
stress due to bending.
Final Design Decisions and Comments
In summary, the following are the design decisions for the horizontal
bar of the conveyor hanger shown in Figure 5–18.
1. Material: SAE 1020 hot-rolled steel.
2. Size: Rectangular cross section. Thickness t = 0.50 in; height h = 2.00
in.
3. Overall design: Figure 5–18 shows the basic features of the bar.
4. Other considerations: Remaining to be specified are the tolerances
on the dimensions for the bar and the finishing of its surfaces.
The potential for corrosion should be considered and may call for
paint, plating, or some other corrosion protection.
The size of the cross section can likely be used with the as-received
tolerances on thickness and height, but this is somewhat dependent
on the design of the fixture that holds the engine block and the
conveyor hangers. So the final tolerances will be left open pending
later design decisions.
The holes in the bar for the pins should be designed to produce a
close sliding fit with the pins, and the details of specifying the
tolerances on the hole diameters for such a fit are discussed in
Chapter 13. See also the discussion on lug joints in Section 3–21.
5. Other possible modes of failure: The analysis used in this problem
assumed that failure would occur due to the bending stresses in the
rectangular bar.
The dimensions were specified to preclude this from happening.
Other possible modes are discussed as follows:
a. Deflection of the bar as an indication of stiffness: The type of
conveyor system described in this problem should not be expected to
have extreme rigidity because moderate deflection of members
should not impair its operation. However, if the horizontal bar deflects
so much that it appears to be rather flexible, it would be deemed
unsuitable. This is a subjective judgment.
 We can use Case (a) in Table A14–1 to compute the deflection.

In this design,
F = 310 lb = maximum load on the bar
L = 24.0 in = distance between supports
E = 30 * 106 psi = modulus of elasticity of steel
I = th3/12 = moment of inertia of the cross section
I = (0.50 in)(2.00 in)3/12 = 0.333 in4
Then,

This value seems satisfactory. In Section 5–11, some guidelines were

given for deflection of machine elements. One stated that bending
Qdeflections for general machine parts should be limited to the range
of 0.0005 to 0.003 in/in of beam length.
For the bar in this design, the ratio of y/L can be compared to this
range:
y/L = (0.0089 in)/(24.0 in) = 0.0004 in/in of beam length
Therefore, this deflection is well within the recommended range.
b. Buckling of the bar: When a beam with a tall, thin, rectangular
cross section is subjected to bending, it would be possible for the
shape to distort due to buckling before the bending stresses would
cause failure of the material. This is called elastic instability, and a
complete discussion is beyond the scope of this book. However,
Reference 14 shows a method of computing the critical buckling load
for this kind of loading. The pertinent geometrical feature is the ratio
of the thickness t of the bar to its height h. It can be shown that the
bar as designed will not buckle.
c. Bearing stress on the inside surfaces of the holes in the beam: Pins
transfer loads between the bar and the mating elements in the
conveyor system.
 It is possible that the bearing stress at the pin–hole interface could be
large, leading to excessive deformation or wear. Reference 4
in Chapter 3 indicates that the allowable bearing stress for a steel pin
in a steel hole is 0.90𝑆𝑌 .

The actual bearing stress at the center hole is found using the
projected area, Dpt.

Thus the pin and hole are very safe for bearing
Design Example 5–4
A bracket is made by welding a rectangular bar to a circular rod, as shown in Figure
5–19. Design the bar and the rod to carry a static load of 250 lb.
Solution
Objective The design process will be divided into two parts:
1. Design the rectangular bar for the bracket.
2. Design the circular rod for the bracket.
Rectangular Bar
Given The bracket design is shown in Figure 5–19. The rectangular bar
carries a load of 250 lb vertically downward at its end.
An effectively fixed support is provided by the weld at its left end
where the loads are transferred to the circular rod.
The bar acts as a cantilever beam, 12 in long. The design task is to
specify the material for the bar and the dimensions of its cross
section.
Basic Design Decisions
We will use steel for both parts of the bracket because of its relatively
high stiffness, the ease of welding, and the wide range of strengths
available.
Let’s specify SAE 1340 annealed steel having 𝜎𝑌 = 63 ksi and
𝜎𝑈 = 102 ksi (Appendix 3). The steel is highly ductile, with a 26%
elongation.
 The objective of the design analysis that follows is to determine the
size of the cross section of the rectangular bar. Assuming that the bar
acts as a cantilever and the loading and processing conditions are well
known, we will use a design factor of N = 2 because of the static load.
Analysis and Results
The free-body diagram of the cantilever bar is shown in Figure 5–20,
along with the shearing force and bending moment diagrams.
This should be a familiar case, leading to the judgment that the
maximum tensile stress occurs at the top of the bar near to where it is
supported by the circular rod.
This point is labeled element A in Figure 5–20. The maximum bending
moment there is M = 3000 lb.in.
The stresses at A are
 where S = section modulus of the cross section of the bar.
We will first compute the minimum allowable value for S and then
determine the dimensions for the cross section.
The MSST criterion, Equation (5–7), applies because of the static
loading. We will first compute the design stress from

Now we must ensure that the expected maximum stress 𝜎𝐴 = 𝑀Τ𝑆

does not exceed the design stress.
We can substitute sA = sd and solve for S.

The relationship for S is

As a design decision, let’s specify the approximate proportion for the

cross-sectional dimensions to be h = 3t. Then,
 The required minimum thickness is then

The nominal height of the cross section should be, approximately,

Final Design Decisions and Comments

In the fractional-inch system, preferred sizes are selected to be t = 3/8
in = 0.375 in and h = 1 1/4 in = 1.25 in (see Table A2–1).
Note that we chose a slightly smaller value for t but a slightly
larger value for h. We must check to see that the resulting value for S
is satisfactory.
 This is larger than the required value of 0.095 in3, so the design is
satisfactory.
Part (b) of Figure 5–20 shows a finite-element analysis (FEA) model for
the rectangular bar only.
The variations in color on the bar represent stress levels with red
being the highest and blue the lowest.
This model verifies the observation that the highest stress is at
Element A.
Circular Rod
Given The bracket design is shown in Figure 5–19. The design task is to
specify the material for the rod and the diameter of its cross section.
Basic Design Decisions
Let’s specify SAE 1340 annealed steel, the same as that used for the
rectangular bar. Its properties are
Analysis and Results
Figure 5–21 is the free-body diagram for the rod. The rod is loaded at
its left end by the reactions at the end of the rectangular bar, namely,
a downward force of 250 lb and a moment of 3000 lb.in.
The figure shows that the moment acts as a torque on the circular rod,
and the 250-lb force causes bending with a maximum bending
moment of 2000 lb # in at the right end. Reactions are provided by the
weld at its right end where the loads are transferred to the support.
The rod then is subjected to a combined stress due to torsion and
bending. Element B on the top of the rod is subjected to the maximum
combined stress.
The manner of loading on the circular rod is identical to that analyzed
earlier. It was shown that when only bending and torsional shear
occur, a procedure called the equivalent torque method can be used
to complete the analysis. First we define the equivalent torque, Te:
Then the shear stress in the bar is

For a solid circular rod,

Our approach is to determine the design shear stress and Te and then solve for Zp.
The maximum shear stress theory of failure can be applied and the design shear
stress is

We let 𝜏 = 𝜏𝑑 and solve for Zp:

 Now that we know Zp, we can compute the required diameter from

This is the minimum acceptable diameter for the rod.

Final Design Decisions and Comments
The circular rod is to be welded to the side of the rectangular bar, and
we have specified the height of the bar to be 1 1/4 in.
Let’s specify the diameter of the circular rod to be machined to 1.10
in. This will allow welding all around its periphery.
Figure 5–21(b) shows a finite-element analysis (FEA) model for the
circular bar only.
The variations in color on the bar represent stress levels with red
being the highest and blue the lowest.
This model verifies the observation that the highest stress is at
Element B.