You are on page 1of 14

Fatigue

1 FLUCTUATING STRESSES
In the early chapters, the external forces acting on a machine component were assumed to be static.
In many applications, the components arc subjected to forces, which are not static, but vary in
magnitude with respect to time. The stresses induced due to such forces are called fluctuating
stresses. It is observed that about 80% of failures of mechanical components are due to fatigue
failure resulting from fluctuating stresses. In practice, the pattern of stress variation is irregular
and unpredictable, as in case of stresses due to vibrations. For the purpose of design analysis,
simple models for stress-time relationships are used.

There are three types of mathematical models for cyclic stresses—fluctuating or alternating
stresses, repeated stresses and reversed stresses.

The fluctuating or alternating stress varies in


a sinusoidal manner with respect to time. It has
some mean value as well as amplitude value. It
fluctuates between two limits—maximum and
minimum stress. The stress can be tensile or
compressive or partly tensile and partly
compressive.

The repeated stress varies in a sinusoidal


manner with respect to time, but the variation
is from zero to some maximum value. The
minimum stress is zero in this case and
therefore, amplitude stress and mean stress arc
equal.
The reversed stress varies in a sinusoidal
manner with respect to time, but it has zero
mean stress. In this case, half portion of the
cycle consists of tensile stress and the
remaining half of compressive stress.

σmax and σmin are maximum and minimum stresses, while σm and σa are called mean stress and
stress amplitude respectively. It can be proved that,

2 FATIGUE FAILURE

It has been observed that materials fail under fluctuating stresses at a stress magnitude which is
lower than the ultimate tensile strength of the material or yield strength of the material. Such type
of failure is known as Fatigue failure. Further, it has been found that the magnitude of the stress
causing fatigue failure decreases as the number of stress cycles increase. This phenomenon of
decreased resistance of the materials to fluctuating stresses is the main characteristic of fatigue
failure.

Fatigue failure is also defined as time delayed fracture under cyclic loading. Examples of parts
in which fatigue failures are common are transmission shafts, connecting rods, gears, vehicle
suspension springs and ball bearings.

Fatigue failure begins with a crack at some point in the material. The crack is more likely to occur
in the following regions:

(i) Regions of discontinuity, such as oil holes, keyways, screw threads, etc.
(ii) Regions of irregularities in machining operations, such as scratches on the surface, stamp
mark, inspection marks, etc.

(iii) Internal cracks due to defects in materials like blow holes

These regions are subjected to stress concentration due to the crack. The crack spreads due to
fluctuating stresses, until the cross-section of the component is so reduced that the remaining
portion is subjected to sudden fracture.

3 ENDURANCE LIMIT

The fatigue or endurance limit of a material is defined as the maximum amplitude of completely
reversed stress that the standard specimen can sustain for an unlimited number of cycles without
fatigue failure. Since the fatigue test cannot be conducted for unlimited or infinite number of cycles,
106 cycles is considered as a sufficient number of cycles to define the endurance limit. There is
another term called fatigue life, which is frequently used with endurance limit. The fatigue life is
defined as the number of stress cycles that the standard specimen can complete during the test
before the appearance of the first fatigue crack.

3.1 FATIGUE TEST

In this test, a carefully machined and highly polished specimen subjected "to a constant, bending
moment is made to rotate at a high speed, thus subjecting it to cycles of stress reversals of
completely reversed nature. On the first specimen, maximum stress is kept slightly lower than its
ultimate strength and the no. of cycles of stress reversals required to cause fracture is observed.
For the next specimen, stress is reduced and no. of rotations to cause fracture is noted down.
Experiments are conducted on several specimen, each time reducing the stress level and finding
out the no. of stress reversal cycles at fracture. S-N diagram is obtained by plotting results of
various trials, on a log-log graph sheet. It is noted that, below a certain stress, failure will not occur,
no matter how great the number of cycles are, as represented by the horizontal portion of the
diagram.
A beam of circular cross-section is subjected to bending moment Mb. Under the action of bending
moment, tensile stresses are induced in the upper half of the beam and compressive stresses in the
lower half. Let us consider point A on surface of the beam and let us try to find out stresses at this
point when the shaft is rotated through one revolution. Initially, the point A occupies position A1
in the central horizontal plane with zero stress. When the shaft is rotated through 90°, it occupies
the position A2. It is subjected to maximum tensile stress in this position. When the shaft is further
rotated through 90°, the point A will occupy the position A3 in the central horizontal plane with
zero stress. A further rotation of 90° will bring the point A to the position A4. It is subjected to
maximum compressive stress in this position. The variation of stresses
at the point A during one revolution of the beam is shown in Figure.

It is observed that the beam is subjected to completely reversed stresses with tensile stress in the
first half and compressive stress in the second half. The distribution is sinusoidal and one stress
cycle is completed in one revolution.

The S-N curve is the graphical representation of stress amplitude versus the number of stress
cycles before the fatigue failure on a log-log graph paper. The S-N curve for steels is illustrated
in Fig. Each test on the fatigue testing machine gives one failure point on the S-N diagram. In
practice, the points are scattered in the figure and an average curve is drawn through them.
Figure: S-N curve for Steel

For ferrous materials like steels, the S-N curve becomes asymptotic at 106 cycles, which indicates
the stress amplitude corresponding to infinite number of stress cycles. The magnitude of this stress
amplitude at 106 cycles represents the endurance limit of the material.

The endurance limit, in a true sense, is not exactly a property of material like ultimate tensile
strength. It is affected by factors such as the size of the component, shape of component, the surface
finish, temperature and the notch sensitivity of the material.

4 LOW CYCLE AND HIGH CYCLE FATIGUE

Fig: Low Cycle Fatigue and High Cycle Fatigue

The complete S-N curve from 100 cycle to 108 cycles is shown in Figure. There are two regions of
this curve namely, low-cycle fatigue and high-cycle fatigue.

The difference between these two fatigue failures is as follows:


(i) Any fatigue failure when the number of stress cycles are less than 1000, is called low-cycle
fatigue. Any fatigue failure when the number of stress cycles are more than 1000, is called high-
cycle fatigue.

(ii) Failure of studs on truck wheels, failure of setscrews for locating gears on shafts or failures
of short-lived devices such as missiles are the examples of low-cycle fatigue. The failure of
machine components such as springs, ball bearings or gears that are subjected to fluctuating
stresses, are the examples of high-cycle fatigue.

(iii) The low-cycle fatigue involves plastic yielding at localized areas of the components.
Components subjected to high-cycle fatigue are designed on the basis of endurance limit stress.

5 NOTCH SENSITIVITY

It is observed that the actual reduction in the endurance limit of a material due to stress
concentration is less than the amount indicated by the theoretical stress concentration factor Kt.

Fatigue stress concentration factor, which is defined as follows:

Endurance limit of the notch free specimen


𝐾𝑡𝑓 =
Endurance limit of the notched specimen

This factor Ktf is applicable to actual materials and depends upon the grain size of the material.

Notch sensitivity is defined as the susceptibility of a material to succumb to the damaging effects
of stress raising notches in fatigue loading.

The notch sensitivity factor q is defined as

Increase of actual stress over nominal stress


𝑞=
Increase of theoretical stress over nominal stress

(𝐾𝑡𝑓 𝜎𝑜 − 𝜎𝑜 )
𝑞=
(𝐾𝑡 𝜎𝑜 − 𝜎𝑜 )

Rearranging it,
𝐾𝑡𝑓 =1+q (𝐾𝑡 − 1)

When q=0, 𝐾𝑡𝑓 = 1, the material has no sensitivity to notches

When q=1, 𝐾𝑡𝑓 = 𝐾𝑡 , the material has full sensitivity to notches

5.1 ENDURANCE LIMIT – FACTORS

The laboratory method for determining the endurance limit of materials, although more precise, is
laborious and time consuming. A number of tests are required to prepare one S-N curve and each
test takes considerable time. It is, therefore, not possible to get the experimental data of each and
every material.

The endurance limit of a particular component (machine member) (𝜎𝑒𝑛1 ) is different from the
endurance limit of a rotating beam specimen (𝜎𝑒𝑛 ) due to a number of factors. The difference
arises due to the fact that there are standard specifications and working condition for the rotating
beam specimen, while the actual components work under different conditions. Different modifying
factors (A, B, C) are used in practice to account for this difference.

𝜎𝑒𝑛1 = 𝜎𝑒𝑛 . 𝐴. 𝐵. 𝐶

𝜎𝑒𝑛1 − 𝐸𝑛𝑑𝑢𝑟𝑎𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡 𝑜𝑓 𝑎 𝑝𝑎𝑟𝑡𝑖𝑐𝑢𝑙𝑎𝑟 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡

𝜎𝑒𝑛 − 𝐸𝑛𝑑𝑢𝑟𝑎𝑛𝑐𝑒 𝑙𝑖𝑚𝑖𝑡 𝑜𝑓 𝑎 𝑟𝑜𝑡𝑎𝑡𝑖𝑛𝑔 𝑏𝑒𝑎𝑚 𝑠𝑝𝑒𝑐𝑖𝑚𝑒𝑛

5.1.1. Loading Factor (A)


The endurance limit is reduced due to stress concentration. When the component is subjected to
an axial fluctuating load, the conditions are different. In axial loading, the entire cross-section is
uniformly stressed to the maximum value. In the rotating beam test, the specimen is subjected to
bending stress. The bending stress is zero at the center and outer region near the surface subjected
to maximum stress.. There is more likelihood of a microcrack being present in the much higher
high-stress field of axial loading than in the smaller volume outer region of the rotating beam
specimen. Therefore, endurance limit in axial loading is lower than the rotating beam test.

A = 1.0 for reversed bending


= 0.7 to 1.0 for reversed axial loading

= 0.5 to 0.6 for reversed torsional loading

5.1.2. Size Correction Factor (B)


The larger the machine part, the greater the probability that a flaw exists somewhere in the
component. Size factor B takes into account the reduction in endurance limit due to increase in the
size of the component.

5.1.3. Surface Finish Factor (C)


The surface of the rotating beam specimen is polished to mirror finish. This makes the specimen
almost free from surface scratches and imperfections. It is impractical to provide such an expensive
surface finish for the actual component. When the surface finish is poor, there are scratches and
geometric irregularities on the surface which results in stress concentration. The endurance limit
is reduced due to introduction of stress concentration at these scratches. The surface finish factor
takes into account the reduction in endurance limit due to the variation in the surface finish between
the specimen and the actual component.

6 SODERBERG LINE AND GOODMAN LINE

The failure points corresponding to simple tension test and different types of fatigue tests are
plotted as shown in figure. The curve obtaining by joining the points is called Gerber’s Parabola.
Straight line obtained by joining the endurance limit to ultimate tensile strength is Goodman line.
Soderberg line is that which joins endurance limit to the yield point. A machine member does not
fail; when it is subjected to a combination mean stress and stress amplitude corresponding to any
point inside the Soderberg line.
6.1 SODERBERG’S DESIGN EQUATION (APPLICABLE FOR DUCTILE MATERIALS)

Consider a point P, representing mean stress of σm and stress amplitude of σa on Soderberg safe
stress line for a factor of safety of N. The triangles AOB and PCB are similar.

𝑃𝐶 𝐵𝐶
=
𝐴𝑂 𝐵𝑂
In terms of stress values,

𝜎𝑦𝑝
𝜎𝑎 (⁄𝑛 − 𝜎𝑚 )
𝜎 = 𝜎
( 𝑒𝑛⁄𝑛) ( 𝑦𝑝⁄𝑛)

𝑛. 𝜎𝑎 (𝑛. 𝜎𝑚 )
= 1−
𝜎𝑒𝑛 (𝜎𝑦𝑝 )

𝜎𝑎 1 𝜎𝑚
= −
𝜎𝑒𝑛 𝑛 𝜎𝑦𝑝

𝜎𝑎 𝜎𝑚 1
+ =
𝜎𝑒𝑛 𝜎𝑦𝑝 𝑛

Factors Kt and Ktf are to be introduced along with σm and σa to take care of stress raisers in the
machine members. 𝜎𝑒𝑛 is to be replaced by 𝜎𝑒𝑛1 for a machine member (particular component).

𝜎𝑎 𝜎𝑚 1
1
+ =
𝜎𝑒𝑛 𝜎𝑦𝑝 𝑛

Thus the Soderberg’s equation becomes

𝐾𝑡𝑓 𝜎𝑎 𝐾𝑡 𝜎𝑚 1
+ =
𝐴𝐵𝐶𝜎𝑒𝑛 𝜎𝑦𝑝 𝑛

Theoretical stress concentration factor (𝐾𝑡 ) is neglected for ductile materials for fatigue loading.

Soderberg Equation becomes,

𝐾𝑡𝑓 𝜎𝑎 𝜎𝑚 1
+ =
𝐴𝐵𝐶𝜎𝑒𝑛 𝜎𝑦𝑝 𝑛
7 NUMERICALS

1. Determine the location and magnitude of the maximum stress in the member shown in
figure 7.1. If the factor of safety required is 2, what grade of steel do you suggest? If the
bending moment on the member is completely reversed, what would be the factor of safety?
2. Determine the maximum load for a simply supported beam of 50 mm diameter and 600
mm span, centrally loaded, as the load cyclically varies from W to 3W. σen = 350 MPa, σyp
= 525 MPa and σu = 700 MPa. Design factor of safety is 1.3.
3. A cold drawn steel shaft subjected to a torque that varies from -12×104 Nmm to 48 ×104
Nmm is shown in the figure 7.3. Determine the size of the shaft. σen = 225 MPa, σy = 300
MPa and σu = 450 MPa. Design factor of safety is 1.5
4. A component with σu = 630 MPa and σen = 315 MPa is shown in the figure 7.4. It is
subjected to a completely reversed axial load of 25 kN. Factor of safety is 2. Take size
factor as 0.85, notch sensitivity factor as 0.8. Determine the plate thickness.
5. A hot rolled steel shaft is subjected to a torsional moment that varies from 330 Nm
(clockwise) to 110 Nm (anticlockwise) as an applied bending moment at the critical section
varies from +440 Nm to -220 Nm. The shaft is of uniform cross section and there is no
keyway at the critical section. Determine the diameter of the shaft. σu = 550 MPa, σy = 410
MPa and n =1.5.
6. A steel cantilever member shown in the Fig.7.5 is subjected to a transverse load at its end
that varies from 50 N(upward) to 150 N (downward) and an axial load that varies from 125
N (in compression) to 500 N ( tension). Determine the required diameter ,at the change of
section for infinite life using a factor of safety of 2.The theoretical stress concentration
factor for bending and axial loads are 1.5 and 1.6 respectively at the stress raiser. The
properties of the material are
σu = 560 MPa, σy = 460 MPa and σen = 280 MPa
Fig: 7.1

Fig: 7.2

Fig: 7.3

Fig: 7.4
Fig: 7.5

You might also like