Materials and Manufacturing

Fracture Toughness and Fatigue

Dr. Ryan McFadden

 Stress Concentrations
In simple analysis it is often assumed that the cross section of components remains uniform or changes gradually. In
practice there can be abrupt changes in cross section, which lead to localised regions of high stress called stress
concentrations (or stress raisers).

Stress concentrations must be considered for fatigue and impact loading of most materials. However, stress-concentration
factors are rarely applied to ductile materials under static loading based on the following assumptions:

1. Areas of high stress caused by stress concentrations are highly localised and will not dictate the performance of the part.

2. If the magnitude of the loading is large enough to cause yielding due to the stress concentration, the localized area will
plastically deform immediately upon loading;

3. Ductile materials typically strain-harden on yielding, resulting in a localized increase in material strength;
 max =  nom
Stress concentrations are a significant concern however in brittle materials, which are susceptible to abrupt failure due to
their inability to yield and plastically deform.
Crack Initiation

• Cracks will normally start from a stress concentration in a brittle material. This could be a geometrical feature,
a notch or hole in the material, or an inclusion.

• The ratio of the raised stress level at a stress concentration to the nominal stress level in the rest of the
material is known as the stress concentration factor, k.

 max
k=
 nom
• The nominal stress is calculated by elementary theory. The maximum stress for axial load, torsion, and
bending are given by;
F Tr My
 max = k  max =k  max = k
A J I

• For more complex situations stress concentrations may be determined experimentally.

Example
 Determination of K Value

1. Experimental Methods 2.Computational Methods

• Optical methods, such as photoelasticity Finite-element techniques
• Brittle-coating method provide a powerful
• Strain gauging. computational method of
assessing stress-
concentration factors
Example – using standard tables and graphs to find k value

• Often for standard geometries, tables and graphs

have been compiled which allow easy
determination of k value. Here is an example for a
flat bar with fillets.
Avoidance of high stress concentrations by using good design

- Increasing radii on corners

- Minimising sudden large changes in
geometry
- Polishing surfaces to remove cracks and
defects caused by processing
- If notches or grooves are required, a
number of small ones are preferable to
a large one
- Sometimes stress relieving grooves or
holes are placed behind a sharp notch Ref:
https://nptel.ac.in/c
ourses/112105125/
pdf/Module-
3_lesson-2.pdf
Fracture Mechanics
• Early fracture mechanics was developed by Griffith, • Where
in order to explain why the observed strength of
materials is considerably less than the theoretical
strength.

• He concluded that real materials contain defects

and cracks which reduce their strength

• These defects cause stress concentrations • This is a measure of a material’s resistance to fast
fracture (or brittle fracture) when a crack is present
• Depends on temperature, microstructure and strain rate
• Fracture toughness is given by the equation: including other factors

• When crack size is small in comparison to the size of

K IC = Y a the component, Y is generally taken as 1
 Determining the fracture toughness
• Fracture toughness KIC can be calculated using long term testing at different loads and temperatures on
specimens which have been notched to a specific geometry.

• The growth of the notch (or pre-crack) is monitored, and using the data in conjunction with the theory of linear
elastic fracture mechanics (LEFM), the KIC can be calculated.

• This process can be very difficult and expensive

• The values obtained through long-term notch testing combined with LEFM also have a strong correlation the with
results of notched impact tests, which are much easier, faster and cheaper to carry out

• Using notch-testing can determine the “notch sensitivity”. Materials that are “notch sensitive” are prone to brittle
failure, as cracks will propagate easily through them if a notch or a small crack is present

• The two test methods are known as the Charpy test and the Izod test
Example 1
Example 1 – Solution
Exam Questions
 Fatigue
• Fatigue is the failure of a component under cyclic loading below the ultimate
tensile strength (UTS) of the material (from a tensile test). This means that
the components/structures are subjected to alternating stresses, e.g.
suspension units on a car, connecting rods or crank shafts, aircraft structures.
• Failure usually occurs in the following manner;

1. Crack initiation (surface dislocations, poor surface finish, notches, other

stress concentrations, accelerates this process)
2. Slow crack propagation (crack spreads through the material)
3. When the critical crack length is reached, the crack propagates very fast to
failure.
Types of fatigue loading and failure
• Fatigue of cracked structures
-cracks pre-exist: propagation controlled fracture. Example almost any large structure, particularly
those containing welds such as bridges, ships and pressure vessels.
• Fatigue of uncracked components
No cracks pre-exist; initiation-controlled fracture. Example, almost any small components such as pins,
ball races. Gear teeth, axles, crank shafts, drive shafts.
• High Cycle
- Stress below yield strength; over 104 cycles to fracture. Rotating or vibrating systems; wheels, axles,
engine components.
- Macroscopically brittle and may be very long life
• Low Cycle
- Stress exceeds yield strength; lower than 104 cycles to fracture-Airframe, turbine, core components
of nuclear reactor or other components subject to occasional overload. (Lots of plastic deformation).
Fracture surface after fatigue failure

• Macroscopic marks – “Beach marks” • Microscopic marks within each beachmark known
– Visible with the naked eye as “Striations”
– Each beach mark represents a period of time – Visible under electron microscope
over which crack growth occurred – Each striation represents the growth of the crack
– E.g. if a machine is run for 3 hours a day, a at single loading cycle
beachmark would start at the beginning of those – Striation width increases as stress increases
3 hours and stop at the end of the 3 hours – There may be many thousands of striations in a
single beachmark
 Jack hammer component,
shows no yielding before
fracture

Catastrophic failure mode!

Crack initiation site

Fracture zone
Propagation zone
 Crank shaft

Gear tooth failure

 Hawaii, Aloha Flight 243, a Boeing 737, an upper part of the plane's cabin area
rips off in mid-flight. Metal fatigue was the cause of the failure.

Pictures Ref: Ken Youssefi

https://www.youtube.com/watch?v=QBOcTX6-OrE

MAE dept., SJSU

S-N Curve

• A plot of stress (S) versus the number of cycles to failure (N) is called an S-N curve.

UTS
Stress
(S)

Fatigue limit SD

Number of cycles to failure

(N)
• The presence of stress concentrations (sharp internal corners, tool marks, holes and notches, for example) can
considerably reduce the fatigue life of the component.
• In ferrous metals, there is a certain stress level below which failure will not occur no matter how many stress
reversals there are. This is called the fatigue limit or endurance limit SD. As a general rule, for steels, the fatigue
limit is about half the UTS of the material.
Non-ferrous metals

• Many non-ferrous materials do not have a fatigue limit and they will eventually fail even at very low levels of
cyclic loading.
•For these metals a
S
“fatigue strength” is
specified

•Fatigue strength is
defined as the stress
level at which failure
Fatigue will occur for a
Strength specified number of
cycles (e.g. 107 cycles)

N
Example
Terminology in fatigue calculations

Remember to take tensile stress as positive and compressive

stress as negative when performing calculations!
Fatigue life is the number of cycles which will cause failure at a
given stress level
https://www.youtube.com/watch?v=LhUclxBUV_E Fatigue test
Types of stress cycle

• Reversed stress cycle – stress alternates from a max

tensile to max compressive stress of equal
magnitude

• Repeated stress cycle – stress alternates between

tensile and compressive, but the values of these are
asymmetrical about the zero stress level

• Random stress cycle – random pattern to the

stresses applied
Problems with fatigue testing

• When fatigue testing specimens there tends to be a

considerable amount of scatter in the data

• Results from fatigue tests are very sensitive to

specimen preparation, surface defects, inherent
variables in the metal, alignment of the specimens
in the test rig etc

• This scatter in the data can lead to a number of

design uncertainties

• Statistical methods are therefore often used to

assess the results of fatigue tests
Factors affecting fatigue life

• We already saw that increasing the stress amplitude reduces the fatigue life

• Increasing the mean stress also decreases fatigue life

• Surface effects – stress concentrations at the surface will decrease fatigue life
– Design factors – stress concentrations
– Surface treatments – shot peening (induces compressive stresses and improves fatigue life)

• Environmental Effects
– Thermal fatigue – normally induced at high temperatures due to fluctuating thermal stresses (dimensional
expansion and contraction with temperature)

– Corrosion fatigue – fatigue that occurs in the presence of cycling mechanical stress and a corrosive
environment
Past exam questions

Complete these two exam questions. Plots should be made on logarithmic graph paper.
Complete the tutorial questions on Handout

1.
2.
Goodman Relationship or Soderberg Relationship - ways to account for a mean stress
which is not zero

• Most fatigue tests are carried out with a mean stress of zero.

• In service this is often not the case. As the mean stress increases the stress amplitude which causes
failure will decrease.

• There are various empirical equations that are used to calculate fatigue strength when the mean stress
is not zero, including the Goodman and Soderberg relationships

• Since perfect alternation (zero mean stress) rarely occurs in practice, S/N curves should not be used
alone without consideration of the mean stress.

• Care must also be taken in using even the Goodman & Soderberg diagrams since they tend to give a
low value of the fatigue limit for ductile materials & high value of fatigue limit for brittle materials.
 Goodman Diagram Soderberg diagram

• In the Goodman diagram the fatigue limit is zero when the • In the Soderberg diagram the fatigue limit is
mean stress is equal to the tensile strength of the material, zero when the mean stress is equal to the yield
since the material will fail at this value before any cyclical stress of the material. Again, the point
loading can commence. (gives the x ordinate)
representing the stress amplitude for a given
number of cycles & the mean stress for the
• For a given number of cycles at a zero mean stress, the material must lie within the shaded area
fatigue limit can be determined. This fatigue limit is bounded by the axes & the ‘Soderberg Line’, if
plotted as the y ordinate
failure by fatigue is to be avoided.
• A straight line is then drawn. Failure is assumed not to
occur if stressed within the black area
Empirical curves to estimate the mean stress effects on fatigue life

a
The Soderberg curve is shown in red (very
conservative) and the Goodman curve in green.
e’
e’ is the effective alternating stress at failure for Nf
cycles.

When the Goodman curve is used , for example, and

the with e’ = e and Nf =  then any combination
yield uts to the left of the line is safe and to the right is unsafe.
Goodman Relationship

stress amplitude

For the Soderberg line, yield strength is substituted for ultimate strength. This is
a very conservative estimate and rarely used.
Past Exam Question