Introduction to fatigue
design to BS EN 1993-1-9
The assessment of fatigue performance is routine in bridge design but is only relevant to specific
elements in buildings which may suffer from fatigue damage. One example of these is crane
runway beams. Richard Henderson of the SCI introduces some of the background.
1 10 100 1000 10000 100000 1000000 10000000
NSC 29
September 18
Introduction
The phenomenon of metal fatigue involves the development of
cracks in elements that are subject to many repeated applications
of loads which are lower than the maximum loads to which the
element is subjected. If fatigue cracks develop unnoticed, they
will eventually result in complete failure of the element with
potentially catastrophic consequences.
History
Research into fatigue in metal structures began as early as
1837 with tests on conveyor chains. A locomotive axle failure
due to fatigue was recognized as the cause of a train accident
at Meudon, near Versailles in 1842. F Braithwaite coined the
term fatigue in his report “On the fatigue and consequent
fracture of metals” published in the ICE minutes of proceedings
in 1854. August Wohler conducted systematic investigations
into metal fatigue of railway axles over a 20 year period from
1852, produced S-N curves illustrating fatigue behaviour and
introduced the idea of an endurance limit. In 1945, A M Miner
developed a design tool based on the Palmgren linear damage
hypothesis. The stress raising effect of small-radius corners and
the consequent effect on fatigue behaviour was established
following investigation into the Comet air disasters of 1953 and
1954.
Basic Concepts
Fatigue cracks usually initiate at a surface defect such as a sharp
corner or a weld toe and develop when subject to fluctuating
stresses above a certain threshold level. The endurance of a
detail or component is the number of cycles to failure under a
fluctuating stress of a constant amplitude. A point can be plotted
on a graph with the number of cycles to failure (N) as abscissa
and the constant amplitude stress (S) as ordinate. Stress range is
defined as the algebraic difference between the two extremes
of a stress cycle so the constant amplitude fluctuating stress
is a constant stress range. By plotting the endurance for each
constant stress range, a curve called an S-N curve can be drawn,
the typical form of which is shown in Figure 1 on a semi-log plot.
The S-N curve exhibits a negative gradient such that a longer
endurance corresponds to a lower stress range. Stresses below
a stress range magnitude called the cut-off limit do not cause
fatigue damage. According to Miner’s rule, fatigue damage can
be summed linearly for a given detail using the S-N curve to
determine the number of cycles to failure Ni for stress range
Δσi. If the detail is subject to a number of cycles ni for the
corresponding stress range, the fatigue damage can summed
for k stress ranges and must be no greater than 1.0. The relevant
expression is:
350
300
250
200
150
100
50
0
Defects in plain steel, welded joints and welded attachments
all affect the fatigue life of a detail. As a result, many fatigue tests
have been carried out on different details to develop S-N curves
that can be used for fatigue damage calculations. Details are
tabulated in BS EN 1993-1-9 (hereinafter denoted EC3-1-9) and
are separated into the following headings.
Table No. Heading
8.1 Plain members and mechanically fastened joints
8.2 Welded built-up sections
8.3 Transverse butt welds
8.4 Weld attachments and stiffeners
8.5 Load carrying welded joints
8.6 Hollow sections (t ≤ 12.5 mm)
8.7 Lattice girder node joints
8.8 Orthotropic decks – closed stringers
8.9 Orthotropic decks – open stringers
8.10 Top flange to web junction of runway beams
Within each table, details are identified and provided with an
identifying number which corresponds to the relevant S-N curve.
The S-N curves for various classes of detail have been idealized
in EC3-1-9 into a set of parallel lines with straight segments,
plotted on a logarithmic scale on both axes and those for direct
stress are shown in Figure 7.1 of the standard. The S-N curves are
identified by a detail category number ΔσC which corresponds to
the reference fatigue strength in MPa for the detail which is equal
to the constant amplitude stress range for an endurance of 2 ×
106 cycles. The curves are shown in Figure 2.
30
Figure 1: Example S-N Curve
Endurance N (cycles)
Stress range S (MPa)
1.0
k
n1
N1 i=1
Technical
/Fatigue_design_of_bridges#The_mechanism_of_fatigue
/Fatigue_design_of_bridges#Relationship_between_stress_range_and_fatigue_endurance
/Welding#Butt_welds
/Stiffeners
/Steel_construction_products#Structural_hollow_sections
/Fatigue_design_of_bridges#Detail_categories