CH 6: Fatigue Failure Resulting from Variable Loading

Some machine elements are subjected to static loads, and for such elements static
failure theories are used to predict failure (yielding or fracture). However, most
machine elements are subjected to varying or fluctuating stresses (due to the
movement) such as shafts, gears, bearings, cams & followers, etc.
Fluctuating stresses (repeated over long period of time) will cause a part to fail
(fracture) at a stress level much smaller than the ultimate strength (or even the yield
strength in some casses).
Unlike static loading where failure usualy can be detected before it happens (due to
the large deflections associated with plastic deformation), fatigue failures are usualy
sudden and therefore dangerous.
Fatigue failure is somehow similar to brittle fracture where the fracture surfaces are
prependicular to the load axis.
 According to Linear-Elastic Fracture Mechanics (LEFM), fatigue failure develops
in three stages:
- Stage1: development of one or more micro cracks
(the size of two to five grains) due to the cyclic local
plastic deformation.
- Stage2: the cracks progress from micro cracks to
larger cracks (macro cracks) and keep growing
making a smooth plateau-like fracture surfaces with
beach marks.
- Stage3: occurs during the final stress cycle where
the remaining material cannot support the load,
thus resulting in a sudden fracture (can be brittle or
ductile fracture).

 Fatigue failure is due to crack

formation and propagation.
Fatigue cracks usually initiate
at locations with high stresses
such as discontinuities (hole,
notch, scratch, sharp corner,
crack, inclusions, etc.).

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 1 of 19
  Fatigue cracks can also initiate at surfaces having rough surface finish or due to
the presence of tensile residual stresses. Thus, all parts subjected to fatigue
loading are heat treated and polished in order to increase the fatigue life.

Fatigue Life Methods

Fatigue failure is a much more complicated phenomenon than static failure where
much complicating factors are involved. Also, testing materials for fatigue
properties is more complicated and much more time consuming than static testing.
Fatigue life methods are aimed to determine the life (number of loading cycles) of
an element until failure.
 There are three major fatigue life methods where each is more accurate for
some types of loading or for some materials. The three methods are: the stress-
life method, the strain-life method, the linear-elastic fracture mechanics
method.
 The fatigue life is usually classified according to the number of loading cycles
into:
 Low cycle fatigue (1≤N≤1000) and for this low number of cycles, designers
sometimes ignore fatigue effects and just use static failure analysis.
 High cycle fatigue (N>103):
 Finite life: from 103 →106 cycles
 Infinite life: more than 106 cycles

The Strain-Life Method

This method relates the fatigue life to the amount of

plastic strain suffered by the part during the repeated
loading cycles.

 When the stress in the material exceeds the yield

strength and the material is plastically deformed,
the material will be strain hardened and the yield
strength will increase (if the part is reloaded
again). However, if the stress direction is reversed
(from tension to compression), the yield strength
in the reversed direction will be smaller than its
initial value. That means that the material has been
softened in the reverse loading direction (this is

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 2 of 19
 referred to as Bauschinger Effect). Each time the stress is reversed, the yield
strength in the other direction is decreased, and the material gets softer and it
undergoes more plastic deformation until fracture occurs.
 The strain-life method is applicable to Low-cycle fatigue.

The Linear Elastic Fracture Mechanics Method

This method assumes that a crack initiates in the material and it keeps growing until
failure occurs (the three stages described above).

 The LEFM approach assumes that a small crack already exists in the material, and
it calculates the number of loading cycles required for the crack to grow to be
large enough to cause the remaining material to fracture completely.
 This method is more applicable to High-cycle fatigue.

The Stress-Life Method

This method relates the fatigue life to the alternating stress level causing failure, but
it does not give any explanation to why fatigue failure happens.

 The stress-life relation is obtained experimentally using Moore high-speed

rotating beam test.
- The test is conducted by subjecting the
rotating beam to a pure bending moment (of
a fixed known magnitude) until failure occurs.
(Due to rotation, the specimen is subjected to
an alternating bending stress)
- The data obtained from the tests is used to generate the fatigue strength vs.
fatigue life diagram which is known as the S-N diagram.
- The first point on the S-N diagram is the ultimate strength which corresponds
to failure in half a cycle.
- The alternating stress amplitude is set to be below the ultimate strength and
the test is run until failure. The stress level and the number of cycles until
failure give a data point on the chart.
- The testing continues and each time the stress amplitude is reduced (such
that the specimen will live longer) and a new point is obtained.

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 3 of 19
 - For steel alloys, the
low-cycle fatigue and
the high-cycle fatigue
(finite and infinite) can
be recognized as having
different slopes. (they
are straight lines, but
keep in mind it is a log-
log curve)
- For steels, if we keep
reducing the stress
amplitude (for each
test), we will reach to a
stress level for which
the specimen will never fail, and this value of stress is known as the
Endurance Limit (Se).
- The number of stress cycles associated with the Endurance Limit defines the
boundary between Finite-life and Infinite-life, and it is usually between 106 to
107 cycles.

 Steel and Titanium alloys have a clear endurance limit, but this is not true for all
materials.
 For instance, Aluminum alloys do not have an endurance limit, and for such
materials the fatigue strength is reported at 5(108) cycles.
 Also, most Polymers do not have an endurance limit.

The Endurance Limit

The determination of the endurance limit is important for designing machine

elements that are subjected to High-cycle fatigue. The common practice when
designing such elements is to make sure that the fatigue stress level in the element
is smaller than the endurance limit of the material being used.

 Finding the Endurance Limit using the rotating beam experiment is time
consuming where it requires testing many samples and the time for each test is
relatively long. Therefore, we try to relate the endurance limit to other

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 4 of 19
 mechanical properties which are easier to find (such as the ultimate tensile
strength).
 The figure shows a plot of the
Endurance Limits versus Tensile
Strengths for a large number of
steel and iron specimens.
- The graph shows a relation
between the ultimate
strength and endurance
limit for ultimate strengths
up to 1400 𝑀𝑃𝑎, then the
endurance limit seems to
have a constant value.
- The relationship between the endurance limit and ultimate strength for steels
is given as:

0.5 𝑆𝑢𝑡 𝑓𝑜𝑟 𝑆𝑢𝑡 ≤ 1400 𝑀𝑃𝑎

𝑆𝑒 ′ = {
700 𝑀𝑃𝑎 𝑓𝑜𝑟 𝑆𝑢𝑡 > 1400 𝑀𝑃𝑎

 The prime (‘) is used to denote that this is the endurance limit value
obtained for the test specimen (modifications are still needed).

Endurance Limit Modifications Factors

The endurance limit is obtained from the rotating beam test. The test is conducted
under closely controlled conditions (polished specimen of small size at a constant
known temperature, etc.). It is not realistic to expect a machine element to have the
exact same endurance limit value as that obtained from the rotating beam test
because it has different conditions (size, surface finish, manufacturing process,
environment, etc.).

 Thus, some modification factors are used to correlate the endurance limit for a
given mechanical element to the value obtained from tests:
Size Temp. Misc.

𝑆𝑒 = 𝑘𝑎 𝑘𝑏 𝑘𝑐 𝑘𝑑 𝑘𝑒 𝑘𝑓 𝑆𝑒 ′

Finish Load Reliability

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 5 of 19
 Where,
𝑆𝑒 : The endurance limit at the critical location of a machine element with the
geometry and conditions of use.

𝑆𝑒 ′ : The endurance limit obtained from the rotating beam test.

𝑘𝑎 … 𝑘𝑓 : Modification factors (obtained experimentally).

Surface Condition Factor (𝑘𝑎 )

The rotating-beam test specimens are highly polished. A rough surface finish will
reduce the endurance limit because there will be a higher potential for crack
initiation.

 The surface condition modification factor depends on the surface finish of the
part (ground, machined, as forged, etc.) and on the tensile strength of the
material. It is given as:
𝑘𝑎 = 𝑎 𝑆𝑢𝑡 𝑏

 Constants 𝑎 & 𝑏 depend on surface condition and are given in Table 6-2.

Size Factor (𝑘𝑏 )

The rotating-beam specimens have a specific (small) diameter (7.6 mm). Parts of
larger size are more likely to contain flaws and to have more non-homogeneities.
 The size factor is given as:
For bending 1.24 𝑑−0.107 2.79 ≤ 𝑑 ≤ 51 𝑚𝑚
𝑘𝑏 = {
and torsion 1.51 𝑑 −0.157 51 < 𝑑 ≤ 254 𝑚𝑚
where 𝑑 is the diameter.
And:
𝑘𝑏 = 1 for axial loading

 When a member with circular cross-section is NOT rotating, an "effective

diameter" value is used instead of the actual diameter, where:
𝑑𝑒 = 0.37 𝑑

 For other non-rotating cross-sections, 𝑑𝑒 is found as follows:

1) Obtain 𝐴0.95𝜎 from Table 6-3
2) Solve eqn. 6-23 (𝐴0.95𝜎 = 0.01046 𝑑2 ) for the “equivalent diameter” 𝑑
3) Find the effective diameter 𝑑𝑒 using the equation above (i.e., eqn. 6-24)

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 6 of 19
Loading Factor (𝑘𝑐 )
The rotating-beam specimen is loaded in bending. Other types of loading will have a
different effect.
 The load factor for the different types of loading is:
1 𝑏𝑒𝑛𝑑𝑖𝑛𝑔
𝑘𝑐 = { 0.85 𝑎𝑥𝑖𝑎𝑙
0.59 𝑡𝑜𝑟𝑠𝑖𝑜𝑛

Temperature Factor (𝑘𝑑 )

When the operating temperature is below room temperature, the material
becomes more brittle. When the temperature is high, the yield strength decreases
and the material becomes more ductile (and creep may occur).
 For steels, the tensile strength, and thus the endurance limit, slightly increases as
temperature rises, then it starts to drop. Thus, the temperature factor is given
as:
𝑘𝑑 = 0.9877 + 0.6507(10−3 )𝑇𝑐 − 0.3414(10−5 )𝑇𝑐2 + 0.5621(10−8 )𝑇𝑐3
− 0.6246(10−11 )𝑇𝑐4

For 40 ≤ 𝑇𝑐 ≤ 540 °𝐶
 The same values calculated by the equation are also given in Table 6-4 where:
𝑆𝑇
𝑘𝑑 = ( )
𝑆𝑅𝑇

Reliability Factor (𝑘𝑒 )

The endurance limit obtained from testing is usually reported at mean value (it has
a normal distribution with 𝜎̂ = 8% ).
 For other values of reliability, 𝑘𝑒 is found from Table 6-5.

Miscellaneous-Effects Factor (𝑘𝑓 )

It is used to account for the reduction of endurance limit due to all other effects
(such as residual stress, corrosion, cyclic frequency, metal spraying, etc.).
However, those effects are not fully characterized and usually not accounted for.
Thus we use (𝑘𝑓 = 1).

Stress Concentration and Notch Sensitivity

Under fatigue loading conditions, crack initiation and growth usually starts in
locations having high stress concentrations (such as grooves, holes, etc.). The

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 7 of 19
presence of stress concentration reduces the fatigue life of an element (and the
endurance limit), and it must be considered in fatigue failure analysis.

However, due to the difference in ductility, the effect of stress concentration on

fatigue properties is not the same for different materials.

 For materials under fatigue loading, the maximum stress near a notch (hole,
fillet, etc.) is:

𝜎𝑚𝑎𝑥 = 𝐾𝑓 𝜎𝑜 or 𝜏𝑚𝑎𝑥 = 𝐾𝑓𝑠 𝜏𝑜

Where,
𝜎𝑜 : is the nominal stress
𝐾𝑓 : is the fatigue stress concentration factor, which is a reduced value of
the stress concentration factor (𝐾𝑡 ) due to the difference in material
sensitivity to the presence of notches.
and 𝐾𝑓 is defined as:
𝑚𝑎𝑥. 𝑠𝑡𝑟𝑒𝑠𝑠 𝑖𝑛 𝑛𝑜𝑡𝑐ℎ𝑒𝑑 𝑠𝑝𝑒𝑐𝑖𝑚𝑒𝑛
𝐾𝑓 =
𝑠𝑡𝑟𝑒𝑠𝑠 𝑖𝑛 𝑛𝑜𝑡𝑐ℎ − 𝑓𝑟𝑒𝑒 𝑠𝑝𝑒𝑐𝑖𝑚𝑒𝑛

 Defining Notch Sensitivity (𝑞) as:

𝐾𝑓 −1 𝐾𝑓𝑠 −1
𝑞= or 𝑞𝑠ℎ𝑒𝑎𝑟 =
𝐾𝑡 −1 𝐾𝑡𝑠 −1

Where the value of 𝑞 ranges from 0 to 1

𝑞 = 0  𝐾𝑓 = 1 (material is not sensitive)
𝑞 = 1  𝐾𝑓 = 𝐾𝑡 (material is fully sensitive)
 Thus,
𝐾𝑓 = 1 + 𝑞(𝐾𝑡 − 1) or 𝐾𝑓𝑠 = 1 + 𝑞𝑠ℎ𝑒𝑎𝑟 (𝐾𝑡𝑠 − 1)

 For Steels and Aluminum (2024), the notch sensitivity for Bending and Axial
loading is found from Figure 6-20, and for Torsion is found from Figure 6-21.

 Alternatively, instead of using the figures, the fatigue stress concentration

factor (𝐾𝑓 ) can be found as:
𝐾𝑡 −1
𝐾𝑓 = 1 + √𝑎 The Neuber equation
1+
√𝑟

where, 𝑟 : radius

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 8 of 19
 √𝑎 : is a material constant known as the Neuber constant.

 For steels, √𝑎 can be found using Eqns. 6-35a & 6-35b given in the text:

(note that 𝑆𝑢𝑡 needs to be in “𝑘𝑝𝑠𝑖” and √𝑎 will be given in “√𝑖𝑛” )

 For cast iron, the notch sensitivity is very low (from 0 to 0.2); but to be
conservative, it is recommended to use 𝑞 = 0.2
 For simple loading, 𝐾𝑓 can be multiplied by the stress value, or the endurance
limit can be reduced by dividing it by 𝐾𝑓 . However, for combined loading, each
type of stress has to be multiplied by its corresponding 𝐾𝑓 value.

Fatigue Strength
In some design applications, the number of load cycles the element is subjected to
is limited (less than 106), and therefore there is no need to design for infinite life
using the endurance limit.
 In such cases, we need to find the Fatigue
Strength associated with the desired life.
 For the High-cycle fatigue (103→106), the line
equation is 𝑆𝑓 = 𝑎𝑁 𝑏 where the constants “𝑎”
(y intercept) and “𝑏” (slope) are determined
from the end points (𝑆𝑓 )103 and (𝑆𝑓 )106 as:
(𝑆𝑓 )2103 𝑙𝑜𝑔(𝜎𝑓′ ⁄𝑆𝑒 )
𝑎= 𝑎𝑛𝑑 𝑏=− 𝑆𝑒 is the modified
𝑆𝑒 𝑙𝑜𝑔(2𝑁𝑒 ) Endurance Limit
Where 𝜎𝑓′ is the "True Stress at Fracture"; and for steels with HB ≤ 500, it is
approximated as:
𝜎𝑓′ = 𝑆𝑢𝑡 + 345 𝑀𝑝𝑎
fraction of
And (𝑆𝑓 )103 can be related to 𝑆𝑢𝑡 as:
(𝑆𝑓 )103 = 𝑓𝑆𝑢𝑡
where 𝑓 is found as:
𝜎𝑓′
𝑓= (2 × 103 )𝑏
𝑆𝑢𝑡

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 9 of 19
  Using the above equations, the value of 𝑓 is found For 𝑆𝑢𝑡 values less than
as a function of 𝑆𝑢𝑡 (using 𝑁𝑒 = 106 ), and it is 490 MPa, use 𝑓 = 0.9
presented in graphical form in Figure 6-18. to be conservative

 If the value of (𝑓) is known, the constant 𝑏 can be directly found as:
1 𝑓𝑆𝑢𝑡
𝑏 = − 𝑙𝑜𝑔 ( )
3 𝑆𝑒
and 𝑎 can be rewritten as:
(𝑓𝑆𝑢𝑡 )2
𝑎=
𝑆𝑒
 Thus, for 103≤ 𝑁 ≤106, the fatigue strength associated with a given life (𝑁) is:
(𝑆𝑓 )𝑁 = 𝑎𝑁 𝑏

and the fatigue life (𝑁) at a given fatigue stress (𝜎) is found as:
1
𝜎 𝑏
𝑁=( )
𝑎

 Studies show that for ductile materials, the Fatigue Stress Concentration Factor
(𝐾𝑓 ) reduces for 𝑁 < 106 , however the conservative approach is to use 𝐾𝑓 as is.

Example: For a rotating-beam specimen made of 1045 CD steel, find:

a) The endurance limit (Ne=106)
b) The fatigue strength corresponding to (5 × 104 ) cycles to failure
c) The expected life under a completely reversed stress of 420 𝑀𝑃𝑎

Solution:
From Table A-20 𝑆𝑢𝑡 = 630 𝑀𝑃𝑎
Note that no modifications are needed
a) 𝑆𝑒 ′ = 0.5(𝑆𝑢𝑡 ) = 315 𝑀𝑃𝑎
since it is a specimen: 𝑆𝑒 = 𝑆𝑒 ′

b) 𝜎𝑓′ = 𝑆𝑢𝑡 + 345 = 975 𝑀𝑃𝑎

𝑙𝑜𝑔(𝜎𝑓′ ⁄𝑆𝑒 ) 𝑙𝑜𝑔(975⁄315)
𝑏=− =− = −0.0779
𝑙𝑜𝑔(2𝑁𝑒 ) 𝑙𝑜𝑔(2 × 106 )
𝜎𝑓′ 975
𝑓= (2 × 103 )𝑏 = (2 × 103 )−0.0779 = 0.856
𝑆𝑢𝑡 630

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 10 of 19
 (𝑓𝑆𝑢𝑡 )2 (0.856 × 630)2
𝑎= = = 923.4 𝑀𝑃𝑎
𝑆𝑒 315

OR, easier, from Figure 6-18: 𝑓 ≅ 0.857

Then,

(𝑓𝑆𝑢𝑡 )2 (0.857 ×630)2

𝑎= = = 925.4 MPa
𝑆𝑒 315

1 𝑓𝑆𝑢𝑡 1 0.857 × 630

𝑏 = − 𝑙𝑜𝑔 ( ) = − 𝑙𝑜𝑔 ( ) = −0.078
3 𝑆𝑒 3 315

(𝑆𝑓 )𝑁 = 𝑎𝑁 𝑏  (𝑆𝑓 )5×104 = 923.4(5 × 104 )−0.0779

 (𝑆𝑓 )5×104 = 397.5 𝑀𝑃𝑎
c)
1 1
𝜎 𝑏 420 −0.0779
𝑁=( ) =( ) = 24.66 × 103 𝑐𝑦𝑐𝑙𝑒𝑠
𝑎 923.4

Example: The two axially loaded bars shown are made of

1050 HR steel and have machined surfaces. The two bars
are subjected to a completely reversed load 𝑃 as shown.
a) Estimate the maximum value of the load 𝑃 for
each of the two bars such that they will have
infinite life (ignore buckling).
b) Find the static and fatigue factors of safety 𝑛𝑠 & 𝑛𝑓
for bar (B) if it is to be subjected to a completely
reversed load of 𝑃 = 50 𝑘𝑁.
c) Estimate the fatigue life of bar (B) under reversed
load of 𝑃 = 150 𝑘𝑁 (use 𝑓 = 0.9)
Solution:
From Table A-20 𝑆𝑢𝑡 = 620 𝑀𝑃𝑎 & 𝑆𝑦 = 340 𝑀𝑃𝑎
a) 𝑆𝑒 ′ = 0.5(𝑆𝑢𝑡 ) = 310 𝑀𝑃𝑎
Modifying factors:
 Surface factor: 𝑘𝑎 = 𝑎 𝑆𝑢𝑡 𝑏 , from Table 6-2: 𝑎 = 4.51, 𝑏 = −0.265
 𝑘𝑎 = 4.51(620)−0.265 = 0.821
- Size factor: 𝑘𝑏 = 1 since the loading is axial
- Loading factor: 𝑘𝑐 = 0.85 (for axial loading)

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 11 of 19
 - Other factors: 𝑘𝑑 = 𝑘𝑒 = 𝑘𝑓 = 1

Stress concentration (for bar B):

From Figure A-15-3 with 𝑤 ⁄𝑑 = 1.2 & 𝑟⁄𝑑 = 0.1  𝐾𝑡 ≈ 2.38

The fatigue stress concentration factor: 𝐾𝑓 = 1 + 𝑞(𝐾𝑡 − 1)
From Figure 6-20 for steel: 𝑞 ≈ 0.81
 𝐾𝑓 = 1 + 0.81(2.38 − 1) = 2.12
Thus the maximum stress for each should not exceed,
Bar (A): 𝑆𝑒 = 𝑘𝑎 𝑘𝑐 𝑆𝑒′ = (0.821)(0.85)(310) = 216.3 𝑀𝑃𝑎
𝑆𝑒 216.3
Bar (B): (𝑆𝑒 )𝑚𝑜𝑑 = = = 102.03 𝑀𝑃𝑎
𝐾𝑓 2.12

And the maximum load 𝑃 for each is,

Bar (A): 𝑃𝑚𝑎𝑥 = 216.3 × (25 × 25) = 135187.5 𝑁
Bar (B): 𝑃𝑚𝑎𝑥 = 102.03 × (25 × 25) = 63767.7 𝑁
 Note that the maximum load for bar (B) is smaller than that of bar (A)
because of the notch.

b) Static factor of safety 𝑛𝑠 :

From Table A-20: 𝜖𝑓 = 0.15  Ductile material, thus stress concentration
is not applicable.

𝑃 50 × 103
𝜎𝑜 = = = 80 𝑀𝑃𝑎
𝐴𝑛𝑒𝑡 25 × 25
𝑆𝑦 340
 𝑛𝑠 = = = 4.25
𝜎𝑜 80

Fatigue factor of safety 𝑛𝑓 :

(𝑆𝑒 )𝑚𝑜𝑑 𝑆𝑒 216.3

𝑛𝑓 = 𝑜𝑟 𝑛𝑓 = = = 1.28
𝜎𝑜 (𝐾𝑓 𝜎𝑜 ) (2.12)(80)

c) If we calculate the fatigue factor of safety with 𝑃 = 150 𝑘𝑁, we will find it to
be less than one, and thus the bar will not live infinite life.
(𝑓𝑆𝑢𝑡 )2 (0.9 × 620)2
𝑎= = = 1439.5 𝑀𝑃𝑎
𝑆𝑒 216.3

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 12 of 19
 1 𝑓𝑆𝑢𝑡 1 0.9 × 620
𝑏 = − 𝑙𝑜𝑔 ( ) = − 𝑙𝑜𝑔 ( ) = −0.137
3 𝑆𝑒 3 216.3
𝑃 150 × 103
𝜎𝑜 = = = 240 𝑀𝑃𝑎
𝐴𝑛𝑒𝑡 25 × 25
𝜎 = 𝐾𝑓 𝜎𝑜 = 2.12 × 240 = 508.8 𝑀𝑃𝑎
1
𝜎 1⁄𝑏 508.8 ⁄−0.137
 𝑁=( ) =( ) = 1.98 × 103 𝑐𝑦𝑐𝑙𝑒𝑠
𝑎 1439.5
 This approach gives more conservative results than dividing (𝑆𝑒 ) by 𝐾𝑓 ,
and using 𝜎𝑜 as is.

Characterizing Fluctuating Stress

In the rotating-beam test, the specimen is subjected

to completely reversed stress cycles (𝜎𝑚𝑎𝑥 = |𝜎𝑚𝑖𝑛 |)

In the case of a rotating shaft subjected to both radial and axial loads (such as with
helical gears), the fluctuating stress pattern will be different since there will be a
component of stress that is always present (due to the axial load).

 The following stress components can be defined

for distinguishing different states of fluctuating
stress:
𝜎𝑚𝑎𝑥 +𝜎𝑚𝑖𝑛
𝜎𝑚 : Mean or average stress, 𝜎𝑚 =
2

𝜎𝑟 : Stress range, 𝜎𝑟 = |𝜎𝑚𝑎𝑥 − 𝜎𝑚𝑖𝑛 |

𝜎𝑎 : Stress amplitude (half of the stress range),

𝜎𝑚𝑎𝑥 −𝜎𝑚𝑖𝑛
𝜎𝑎 = | |
2

 For uniform periodic fluctuating stress, 𝜎𝑚 & 𝜎𝑎 are used to characterize the
stress pattern.

 We also define:
 Stress ratio: 𝑅 = 𝜎𝑚𝑖𝑛 ⁄𝜎𝑚𝑎𝑥
 Amplitude ratio: 𝐴 = 𝜎𝑎 ⁄𝜎𝑚

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 13 of 19
 Some common types of fluctuating stress:
Completely reversed stress:
𝜎𝑚 = 0
𝜎𝑎 = 𝜎𝑚𝑎𝑥 = |𝜎𝑚𝑖𝑛 |

Repeated stress:
Tension 𝜎𝑎 = 𝜎𝑚 = 𝜎𝑚𝑎𝑥 /2
Compression 𝜎𝑚 = 𝜎𝑚𝑖𝑛 /2

General fluctuating stress:

(non-zero mean)
𝜎𝑎 ≠ 𝜎𝑚 ≠ 0

Fatigue Failure Criteria for Fluctuating Stress

When a machine element is subjected to completely reversed stress (zero mean,

𝜎𝑚 = 0), the endurance limit is obtained from the rotating-beam test (after
applying the necessary modifying factors).

However, when the mean (or midrange) is non-zero, the situation is different and a
fatigue failure criteria is needed.

 If we plot the alternating stress component (i.e., stress amplitude) (𝜎𝑎 ) vs. the
mean stress component (𝜎𝑚 ), this will help in distinguishing the different
fluctuating stress scenarios.

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 14 of 19
  When 𝜎𝑚 = 0 & 𝜎𝑎 ≠ 0 , this will be a completely reversed fluctuating stress.
 When 𝜎𝑎 = 0 & 𝜎𝑚 ≠ 0 , this will be a static stress.
 Any combination of 𝜎𝑚 & 𝜎𝑎 will fall between the two extremes (completely
reversed & static).

 Different theories are proposed to predict failure in such cases:

Yield (Langer) line: It connects 𝑆𝑦 on the 𝜎𝑎 axis with 𝑆𝑦 on 𝜎𝑚 axis. But it is not
realistic because 𝑆𝑦 is usually larger than 𝑆𝑒 .

Soderberg line: The most conservative, it connects 𝑆𝑒 on 𝜎𝑎 axis with 𝑆𝑦 on 𝜎𝑚

axis.
𝜎𝑎 𝜎𝑚 1 Where ( 𝑛) is the design factor
+ =
𝑆𝑒 𝑆𝑦 𝑛

ASME-elliptic line: Same as Soderberg, but it uses an ellipse instead of the

straight line.
2
𝑛𝜎𝑎 2 𝑛𝜎𝑚 It fits experimental data better (see fig 6-25)
( ) +( ) =1
𝑆𝑒 𝑆𝑦

Goodman line: It considers failure due to static loading to be at 𝑆𝑢𝑡 rather than
𝑆𝑦 , thus it connects 𝑆𝑒 on 𝜎𝑎 axis with 𝑆𝑢𝑡 on 𝜎𝑚 axis using a straight line.

𝜎𝑎 𝜎𝑚 1
+ =
𝑆𝑒 𝑆𝑢𝑡 𝑛

Gerber line: Same as Goodman but it uses a parabola instead of the straight line.

𝑛𝜎𝑎 𝑛𝜎𝑚 2
+( ) =1
𝑆𝑒 𝑆𝑢𝑡

 The factor of safety is found as:

 It should be noted that 𝑆𝑒 is the modified endurance limit.

 The fatigue stress concentration factor (𝐾𝑓 ) should be multiplied with both 𝜎𝑎 &
𝜎𝑚 for conservative results.

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 15 of 19
 The load line represents any combination of 𝜎𝑎 and 𝜎𝑚 . The intersection of the
load line with any of the failure lines gives the limiting values 𝑆𝑎 and 𝑆𝑚
according to the line it intercepts.

Modified Goodman (Goodman and Langer)

It combines the Goodman and Langer lines.

 The slope of the loading line passing through

the intersection point of the two lines is
called the critical slope, and it is found as:
𝑟𝑐𝑟𝑖𝑡 = 𝑆𝑎 ⁄𝑆𝑚
(𝑆𝑦 − 𝑆𝑒 )𝑆𝑢𝑡
𝑤ℎ𝑒𝑟𝑒 𝑆𝑚 = & 𝑆𝑎 = 𝑆𝑦 − 𝑆𝑚
𝑆𝑢𝑡 − 𝑆𝑒
 According to the slope of the load line (𝑟 = 𝜎𝑎 ⁄𝜎𝑚 ), it could intersect any of the
two lines:
𝑟𝑆𝑒 𝑆𝑢𝑡 𝑆𝑎
𝑟 > 𝑟𝑐𝑟𝑖𝑡  1 𝑆𝑎 = & 𝑆𝑚 = ,
𝑟𝑆𝑢𝑡 + 𝑆𝑒 𝑟
𝑆𝑎 𝑆𝑚 1
𝑛𝑓 = = =𝜎 𝜎
𝜎𝑎 𝜎𝑚 𝑎
+ 𝑚
𝑆𝑒 𝑆𝑢𝑡
𝑟𝑆𝑦 𝑆𝑦
𝑟 < 𝑟𝑐𝑟𝑖𝑡  2 𝑆𝑎 = & 𝑆𝑚 = ,
1+𝑟 1+𝑟
𝑆𝑎 𝑆𝑚 𝑆𝑦
=𝑛𝑠 = =
𝜎𝑎 𝜎𝑚 𝜎𝑎 + 𝜎𝑚
 Where case 2 is considered to be a static (yielding) failure.

 If we plot the Modified Goodman

on stress (𝜎) vs. mean stress (𝜎𝑚 )
axes, we obtain the complete
Modified Goodman diagram where
it defines a failure envelope such
that any alternating stress that
falls inside the diagram will not
cause failure.
The “Complete”
Modified Goodman
Diagram

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 16 of 19
 Also, there are other modified criteria:
 Gerber-Langer (see Table 6-7)
 ASME-elliptic-Langer (see Table 6-8)

Example: A 40 𝑚𝑚 diameter bar has been machined from AISI-1045 CD rod. The
bar is subjected to a fluctuating tensile load varying from 0 to 100 𝑘𝑁. Because of
the ends fillet radius, 𝐾𝑓 = 1.85 is to be used.
Find the critical mean and alternating stress values 𝑆𝑎 & 𝑆𝑚 and the fatigue factor of
safety 𝑛𝑓 according to the Modified Goodman fatigue criterion.
Solution:
From Table A-20 𝑆𝑢𝑡 = 630 𝑀𝑃𝑎 & 𝑆𝑦 = 530 𝑀𝑃𝑎
𝑆𝑒 ′ = 0.5(𝑆𝑢𝑡 ) = 315 𝑀𝑃𝑎
Modifying factors:
 Surface factor: 𝑘𝑎 = 4.51(630)−0.265 = 0.817 (Table 6-2)
 Size factor: 𝑘𝑏 = 1 since the loading is axial
- Loading factor: 𝑘𝑐 = 0.85 (for axial loading)
- Other factors: 𝑘𝑑 = 𝑘𝑒 = 𝑘𝑓 = 1

 𝑆𝑒 = 𝑘𝑎 𝑘𝑐 𝑆𝑒′ = (0.817)(0.85)(315) = 218.8 𝑀𝑃𝑎

𝐹 𝜋
Fluctuating stress: 𝜎 = , 𝐴 = 𝑑 2 = 1257 𝑚𝑚2
𝐴 4
3
100 × 10
𝜎𝑚𝑎𝑥 = = 79.6 𝑀𝑃𝑎 & 𝜎𝑚𝑖𝑛 = 0
1257
𝜎𝑚𝑎𝑥 + 𝜎𝑚𝑖𝑛 𝜎𝑚𝑎𝑥 − 𝜎𝑚𝑖𝑛
𝜎𝑚 𝑜 = = 39.8 𝑀𝑃𝑎 & 𝜎𝑎 𝑜 = = 39.8 𝑀𝑃𝑎
2 2
Applying 𝐾𝑓 to both components: 𝜎𝑚 = 𝐾𝑓 𝜎𝑚 𝑜 & 𝜎𝑎 = 𝐾𝑓 𝜎𝑎 𝑜

 𝜎𝑚 = 𝜎𝑎 = 1.85(39.8) = 73.6 𝑀𝑃𝑎

 The plot shows that the load line

intersects the Goodman line:
𝑟𝑆𝑒 𝑆𝑢𝑡 1(218.8)(620)
𝑆𝑎 = =
𝑟𝑆𝑢𝑡 + 𝑆𝑒 1(620) + 218.8
 𝑆𝑎 = 162.4 𝑀𝑃𝑎 = 𝑆𝑚

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 17 of 19
 1 𝑆𝑎 𝑆𝑚
𝑛𝑓 = 𝜎 𝜎 𝑜𝑟 𝑛𝑓 = =
𝑎
+ 𝑚 𝜎𝑎 𝜎𝑚
𝑆𝑒 𝑆𝑢𝑡
162.4
 𝑛𝑓 = = 2.21
73.6

See Example 6-12 from text

Torsional Fatigue Loading

For shafts that are subjected to fluctuating shear stress with non-zero mean (due to
pulsating torque), a fatigue criterion (ASME elliptic, Gerber, etc.) needs to be used.

 It should be noted that the endurance limit 𝑆𝑒 already accounts for the torsional
loading since 𝑘𝑐 = 0.59 is used in such case.
 Similarly, the yield or ultimate strengths need to be corrected where the “shear
yield strength” 𝑆𝑦𝑠 or the “shear ultimate strength” 𝑆𝑢𝑠 need to be used and
those are found as:
𝑆𝑦𝑠 = 0.577𝑆𝑦 𝑎𝑛𝑑 𝑆𝑢𝑠 = 0.67𝑆𝑢𝑡

Combination of Loading Modes

The procedures presented earlier can be used for fatigue calculations for a
component subjected to general fluctuating stress (or fully reversed stress, which is
easier) under one of the three modes of loading; Axial, Bending or Torsion.

 For a component subjected to general fluctuating stress under combination of

loading modes:

 The stress corresponding to each mode of loading is split into its alternating
(𝜎𝑎 ) and midrange (𝜎𝑚 ) components.

 The fatigue stress concentration factor corresponding to each mode of

loading is applied to the (𝜎𝑎 & 𝜎𝑚 ) of that mode.

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 18 of 19
  An equivalent von Mises stress is calculated for the alternating and midrange
components as:

2
(𝜎𝑎 )𝑎𝑥𝑖𝑎𝑙 2
𝜎𝑎′ = √[(𝐾𝑓 ) (𝜎𝑎 )𝑏𝑒𝑛𝑑𝑖𝑛𝑔 + (𝐾𝑓 ) ] + 3 [(𝐾𝑓𝑠 ) (𝜏𝑎 )𝑡𝑜𝑟𝑠𝑖𝑜𝑛 ]
𝑏𝑒𝑛𝑑𝑖𝑛𝑔 𝑎𝑥𝑖𝑎𝑙 0.85 𝑡𝑜𝑟𝑠𝑖𝑜𝑛

2 2
′
𝜎𝑚 = √[(𝐾𝑓 ) (𝜎𝑚 )𝑏𝑒𝑛𝑑𝑖𝑛𝑔 + (𝐾𝑓 ) (𝜎𝑚 )𝑎𝑥𝑖𝑎𝑙 ] + 3 [(𝐾𝑓𝑠 ) (𝜏𝑚 )𝑡𝑜𝑟𝑠𝑖𝑜𝑛 ]
𝑏𝑒𝑛𝑑𝑖𝑛𝑔 𝑎𝑥𝑖𝑎𝑙 𝑡𝑜𝑟𝑠𝑖𝑜𝑛

 Note that the alternating component of the axial load is divided by 0.85
(i.e., 𝑘𝑐 for axial loading).
 The torsional alternating stress is not divided by its corresponding 𝑘𝑐 value
(i.e., 0.59) since that effect is already accounted for in the von Mises
stress.
 The endurance limit is calculated assuming that the loading is bending (i.e.,
𝑘𝑐 = 1).
 Finally, a fatigue failure criterion (Gerber, Goodman, ASME-elliptic, etc.) is
selected and applied as usual.

A road map summarizing all the important equations for the stress-life
method is given in Sec. 6-17 in the textbook.

Shigley’s Mechanical Engineering Design, 10th Ed. Class Notes by: Dr. Ala Hijazi

CH 6 (R2) Page 19 of 19