You are on page 1of 4

FFT_1

Fatigue failure theories and Design against fluctuating load


The term ‘fatigue’ was first used by Poncelet in 1839 on studying the failure of ductile
material under time varying loads. The mechanism of failure was not yet understood, and the
brittle appearance of the failure surface in a ductile material caused speculation that the
material had somehow become “tired” and embrittled from the load oscillations.

The mechanism of fatigue failure

Fatigue failures always begin at a crack. Either it is already there since manufacturing or it
may have developed over time due to cyclic stressing around stress concentration.

It is important/critical that dynamically loaded parts are designed with minimum stress
concentrations. There are three stages of fatigue failure.

(i) Crack Initiation


(ii) Crack Propagation
(iii) Sudden fracture due to unstable crack growth.

The details of these stages may be referred in fracture mechanics.

Time varying stresses

Any loads that vary with time can potentially cause fatigue failure. In rotating machinery, the
loads tend to be consistent in amplitude over time and repeat with some frequency. In service
equipment (vehicles of all types), the loads tend to be quite variable in amplitude and
frequency over time and may even be random in nature. The shape of the waveform of the
load-time function seems not to have any significant effect on fatigue failure in the absence
of corrosion, so we usually depict the function schematically as a The typical stress-time
functions experienced by rotating machinery modelled as sine waves are shown in Fig. 1.

(i) Fully reversed stress (Fig. 1a). The mean value is zero in this case.
(ii) Repeated stress (Fig. 1b). In this case, the waveform varies from zero to a
maximum with a mean value equal to the alternating component.
(iii) Fluctuating stress (Fig. 1c). It is a general case with all components having non-
zero values.

Fig.1: Alternating, Mean, and Range Values for Fully Reversed, Repeated, and Fluctuating
Cyclic Stresses.
The stress range  is defined as

1
FFT_1

 = max - min (1a)

The alternating component a is found from

 max   min
a  (1b)
2

and the mean component m is

 max   min
m  (1c)
2

Two ratios can be found:

 min a
R A (1d)
 max m

where R is stress ratio and A is amplitude ratio.

When the stress is fully reversed, R = -1 and A = . When the stress is repeated, R = 0 and A
= -1. When the maximum and minimum stresses have the same sign, both R and A are
positive and 0  R 1. These load patterns may result from bending, axial, torsional, or a
combination of these types of stresses. The combination of the bending and torsional stresses
may be experienced by the shaft carrying only spur gears for power transmission while the
combination of the bending, axial and torsional stresses may be experienced by the shaft
carrying helical or bevel gears. We will see that the presence of a mean-stress component can
have a significant effect on the fatigue life.

Fatigue Regimes

Based on the number of stress or strain cycles that the part is expected to undergo in its
lifetime, it is relegated to either a low-cycle fatigue (LCF) regime or a high-cycle fatigue
(HCF) regime. There is no sharp dividing line between the two regimes. HCF starts normally
from 102 to 104 cycles. 103 cycles is a reasonable approximation of the divide between LCF
and HCF.

Fatigue failure models

There are three fatigue failure models in current use and each has a place and a purpose. They
are
(i) the stress-life (S-N) approach
(ii) the strain-life (ε-N) approach
(iii) linearelastic fracture-mechanics (LEFM) approach.

The Stress-Life Approach

2
FFT_1

This is the oldest of the three models and is the most often used for high-cycle fatigue
(HCF) applications where the assembly is expected to last for more than about 103 cycles of
stress. It works best when the load amplitudes are predictable and consistent over the life of
the part. It is a stress-based model, which seeks to determine a fatigue strength and/or an
endurance limit for the material so that the cyclic stresses can be kept below that level and
failure avoided for the required number of cycles. The part is then designed based on the
material’s fatigue strength (or endurance limit) and a safety factor. In effect, this approach
attempts to keep local stresses in notches so low that the crack initiation stage never begins.
The assumption (and design goal) is that stresses and strains everywhere remain in the elastic
region and no local yielding occurs to initiate a crack. This approach is fairly easy to
implement, and large amounts of relevant strength data are available due to its longtime use.
However, it is the most empirical and least accurate of the three models in terms of defining
the true local stress/strain states in the part, especially for low-cycle fatigue (LCF) finite-life
situations where the total number of cycles is expected to be less than about 103 and the
stresses will be high enough to cause local yielding. On the other hand, with certain materials,
the stress-life approach allows the design of parts for infinite life under cyclic loading.

Examples for HCF

The large class of rotating machinery (stationary or mobile) is well served by the
stress-life (S-N) model because the required lives are usually in the HCF range. For example,
consider the number of load cycles (revolutions) required of an automobile-engine crankshaft
over its useful life. Assume a desired 100 000 miles life with no failure of the crank shaft. For
a car tire radius of 1 ft, final drive ratio of about 3:1 and working mostly on top gear, the
number of cycles in life time would be about 2.5x108 cycles. In another example of
automated production machine (page 327 Norton) the number of cycles for drive shaft, gears
and cams would be 1.25x108 cycles for one year life.

The Strain-Life Approach

Because the initiation of a crack involves yielding, a stress-based approach cannot


adequately model this stage of the process. A strain-based model gives a reasonably
accurate picture of the crack initiation stage. It can also account for cumulative damage due
to variations in the cyclic load over the life of the part, such as overloads that may introduce
favorable or unfavorable residual stresses to the failure zone. Combinations of fatigue loading
and high temperature are better handled by this method, because the creep effects can be
included. This method is most often applied to LCF, finite-life problems where the cyclic
stresses are high enough to cause local yielding. It is the most complicated of the three
models to use and requires a computer solution. Test data are still being developed on the
cyclic-strain behavior of various engineering materials.

The LEFM Approach

Fracture-mechanics theory provides the best model of the crack propagation stage of the
process. This method is applied to LCF, finite-life problems where the cyclic stresses are
known to be high enough to cause the formation of cracks and is most useful in predicting the
remaining life of cracked parts in service. It is often used in conjunction with nondestructive
testing (NDT) in a periodic service-inspection program, especially in the aircraft/aerospace
industry. In the absence of a detectable crack, one approach is to assume that a crack smaller

3
FFT_1

than the smallest detectable crack already exists in order to begin the calculation. It gives
more accurate results when a detectable and measurable crack already exists.

Examples for LCF

One class of machinery that typically sees low-cycle fatigue (LCF) is that of
transportation (service) machinery. The airframe of an airplane, the hull of a ship, and the
chassis of a land vehicle see a load-time history that can be quite variable due to storms,
gusts/waves, hard landings/dockings, etc. (for the aircraft/ship) and overloads, potholes, etc.
(for the land craft). Another example of the use of ε-N and LEFM models is in the design and
analysis of gas-turbine rotor blades, which operate under high stresses at high temperatures
and go through LCF thermal cycles at start-up and shutdown.

As we have seen that the shaft and axle design is in high cycle fatigue regime. We will
concentrate on HCF. For designing the machine elements in this regime one should know the
corrected fatigue strength Sf or corrected endurance limit Se.

Answer the following questions for more understanding.

Q1. Why at a microscopic scale, metals are not homogeneous and isotropic?

Q2. Define the term ‘notch’

Q3. What do you understand by stress concentration?

Q4. Differentiate between fatigue strength and endurance limit

You might also like