2.

Fatigue Strength of Welded

Structural Components

Contents
1. Introduction
2. Fatigue Design Curve
3. Strength Categories of Joints and Their Basic
Allowable Stress Ranges
4. Correction Factor for Basic Allowable Stress
Ranges
5. Fatigue Assessment
6. Fatigue Design Curves from IIW and AASHTO

1
Fatigue Design Recommendations
Safety Assessment
in the Fatigue Limit State of Steel Structural Members
Fatigue Design Recommendations
For Steel Structures
By Japanese Society of Steel Construction (JSSC)
1993(in Japanese), 1995(in English)
The Reference of Today’s Lecture

Other Organizations with Recommendations:

AASHTO (American Association of State Highway and Transportation Officials),
IIW (International Institute of Welding) , …

Introduction

2
 Used Steels
in Akashi Bridge (1998)

Stress – Strain Relationship of

Various Types of Steels
Stress

Strain

3
Strength of Various Types of Steels

t=20mm MPa

σa σy σt Price(¥) σt/2.2 σy/1.7

SM400 140 235 400 96,000 182 138
SM490Y 210 355 490 104,500 223 209
SM570 255 450 570 138,500 259 265
HT690 310 590 690 314 347
HT780 350 690 780 355 406

Steels
Steels intended for the Recommendations
are carbon steel and low alloy steel

Ultimate Strengths:
Steels 330MPa-1GPa
Wires up to 1.6GPa
High Strength Bolts up to 1.2GPa

4
 What’s fatigue

retrofit 9

Brittle fracture from fatigue crack

Hoan-bridge in U.S.A

http://www.jsonline.com

5
 Hoan-bridge

I-Girder Web Gusset

http://www.lichtensteinengineers.com/Hoan/Hoan-Failure-Investigation.PDFより引用

The I-95 Brandywine Bridge

Located in New Castle County,
DE over the Brandywine Creek

6
 Crack initiation point is a
lack of fusion in horizontal
crack stiffener

girder web

Horizontal
Stiffener

lower flange

From “FORENSIC ANALYSIS OF THE STEEL GIRDER FRACTURE IN THE I-95 BRANDYWINE RIVER
BRIDGE，Spencer Quiel， University of Notre Dame，August 8, 2003 “

What’s Fatigue
„ Fatigue: Deterioration of a component caused by
crack initiation and/or by the growth of cracks.
„ Crack is initiated and propagate and causes fatigue
failure of the component under the repetition of
load.
„ Fatigue limit : Fatigue strength under constant
amplitude loading corresponding to a high number
of cycles large enough to be considered infinite by
a design code.
IIW Fatigue Recommendation 2005 Feb.

7
 Propagation of Fatigue Crack

Fatigue Length a
Crystal
slip surface crack

striation mix of ductile & brittle

fracture surface
1st stage 2nd stage
ａｉ
Final Break

repetition N

Image View of propagation speed

Fatigue Propagation

Propagation Speed
ΔKth 2 MPa m

C 2.70E-11
m 2.75
da
dN
(
= C ΔK m − ΔK th
m
)
Δσ 50 MPa ：JSSC Fatigue Design
mm/0.1mil.cyc
a(mm) ΔK da/dN(m)
le
1 2.802496 2.77683E-10 0.02776832
5 6.266571 4.01785E-09 0.401785208
10 8.862269 1.07105E-08 1.071048471
20 12.53314 2.8069E-08 2.806902616

8
 Beach Mark

beach marks are inserted with the same interval

example : A-B-A-B-... cyclic loading
A:P=50kN - 800 kN & N=100,000 ,
B:P=400kN-800kN & N=20,000,

Striation

Characteristic fracture surface pattern of

Fatigue crack (magnify ratio:8000)

9
 Why fatigue becomes problems
After the propagation, crack may lead to brittle
fracture to cause structural failure.
Common sense for statistic design is not
applicable. Local stress concentration dominate
the phenomena. Fatigue strength of each
connection type differs.
Evaluation of fatigue damaged part based on
stress analysis is sometimes hard to apply.

Knowledge and experience of fatigue is

required for judgment of retrofitting

Characteristics of fatigue damage 1

• Stress range and repetition

Most influential factors on fatigue are stress range
and number of repetition of it. Fatigue strength of
steel itself is increased with their strength, but fatigue
strength of weld joints has little dependency on the
material strength.

• Fatigue strength : Fatigue crack is initiated from weld

defects,
defects, notch, or stress concentrated part in weld
joints and propagate. Residual stress influences on
the fatigue strength.

10
Characteristics of fatigue damage 2

• Crack propagation
After propagation, Crack will break the material with
brittle fracture,
fracture, or will stop propagation after the
release of stress due to the cracking.

• Relation between stress range and fatigue life

Linear relation on Log-
Log-Log plot

• Fatigue in highway bridge : People considered that

Fatigue of primary member never happen except
Steel deck under vehicle’
vehicle’s load in Japan before.

Fatigue strength
„ Static strength : Yielding 1 time 10 time
and breaking with ductile
manner
„ Fatigue strength : Under
the load repetition, break
without ductility
stressσ Static loading and fatigue loading
Broken
stress range : larege

Not broken
stress range : small
Stress range

Stress range
⊿σ
1 cyclw (Number N )
Fatigue Limit
time

N
S-N Curve

11
Predominant Factors
Controlling Fatigue Strength

1. Joint Types

2. The Magnitude of The Nominal

Stress Range

3. Number of Stress Cycles

Joint Types
1. Welded Connections
o Transverse butt welded joints
o Longitudinal welded joints
o Cruciform joints
o Gusset joints
o Other welded joints

2. Cable Connections

3. High Strength Bolted Connections

12
 Fatigue : initiate from welding
In bridge structures, fatigues are initiated
from stress concentrated part of welding joints.
Strongly dependent on geometry of welding
toe crack toe crack
Root crack

Partial penetration weld Full penetration weld

Fatigue of welded structures
discussed in Lec #6

fatigue strength of weld joint in plate girder

out-of-plane gusset
with fillet weld (l>100)

Base metal
machine finished

non-load-carry cruciform
Full Pene.from joint without bead treatment
Fillet Weld both side

Fatigue strengths are specified for each joint type

13
 Nominal Stress Ranges
Structures subjected to loads
Nominal Stress Distribution at
Section
compression

Longitudinal joint
Bending
Out-of-Plane Gussets

Longitudinal joint

tension
In-Plane Gusset

Fatigue stress on a gusset

„ stress
1. nominal
stress
2. structural hot
spot stress
3. notch stress
concentration
due to weld
bead

14
 Type of Stress for Fatigue Assessment
Type Stress raisers Stress determined Assessment procedure

General analysis of not applicable for

sectional forces using beam Gross average stress fatigue analysis, only Beam model
A theory, from sectional forces component testing
no stress raiser considered
A + macrogeometrical Range of nominal
effects due to the design of stress (also modified Nominal stress
B the component, but or local nominal approach Shell model
excluding stress risers due stress)
to the welded joint itself.
A + B + structural
discontinuities due to the
Range of hot-
hot-spot hot-
hot-spot structural Shell or Solid
structural detail of the
C structural stress stress approach model
welded joint, but excluding
the notch effect of the weld
toe transition
A + B + C + notch stress a) Fracture mechanics Solid detailed
Range of elastic
concentration due to the approach FEM model
notch stress (total
D weld bead notches b) effective notch with bead
stress)
a) actual notch stress stress approach
b) effective notch stress

Type of stress
Local nominal stress
includes
• The effects of macro-
geometric features of
the component
• stress fields in the
vicinity of
concentrated loads
• significant shell
bending stress

15
 modified nominal stress

shear lag Large opening

near concentrated load eccentric joint

Calculation of Nominal Stress

1. use elementary theories of structural mechanics (linear
elastic)
2. FEM may be used in case,
1. over-determined structures
2. macro-geometric discontinuities
• meshing can be simple and coarse (3*t),
• CARE must be taken to ensure that all stress
raising effects of the structural detail of the weld
joints are excluded (stress concentration due to
weld joint)

16
Stress Cycles
Constant Amplitude Stresses
max
Stresses

min

Number of cycles

Variable Amplitude Stresses

max
Stresses

min

Number of cycles

Fatigue Design Curves

17
Typical Fatigue Design Curves
The curve that represents the relationship between
the stress range and the fatigue life
Log-log relationship What N can you get
if stress is reduced
Stress Range, Sr

to half?

Ｎ・⊿σ３=constant

Fatigue
Cut-off Limit

105 106 107 108

Number of Stress Cycles, N

Fatigue Design Curves (1)

Welded Joints Subjected to Normal Stresses

2 x 106 cycles joint class

2 million allowable
fatigue stress range

18
Fatigue Design Curves (2)
Cables and High Strength Bolts Subjected to
Normal Stresses

2 x 106 cycles

K1 to K5:
Strength
Category

Fatigue Design Curves (3)

Welded Joints Subjected to Shear Stresses

2 x 106 cycles

19
 Strength Categories of Joints
And Their Basic Allowable Stress Ranges

Non-Welded Joints (1)

Strength categories
(basic allowable stress ranges)

A
(190)
B
1. Plates
(155)
C
(125)
B (155)
2. Shaped
B
steel
(155)
C
(125)

20
 Non-Welded Joints (2)
3. Seamless tubes B (155)
4. Base plates with circular holes C (125)
B (155)
5. Base
plates with C (125)
cut out C (125)
gussets
D (100)
B (155)
6. Base plates of
friction type bolted C (125)
connection D (100)
7. Base plates of bearing type bolted connection B (155)
8. Base plates with holes and bolts, which do B
not transfer the loads along the direction of (155)
stress

Transverse Butt Welded Joints

1. With ground flush surfaces B(155)

2. With finished weld toe C(125)
D(100)
3. As-welded
joint
D(100)
F (65)
F (65)

21
 Longitudinal Welded Joints

1. Complete penetration B(155)

groove welded joints
from both sides C(125)
2. Partial penetration groove welded joints D(100)
3. Fillet welded joints D(100)
4. Welded joints with backing bars E(80)
5. Intermittent fillet welded joints E(80)
6. Welded joints with copes G(50)
D
7. Welded joints (100)
adjacent to fillets of
E
cut out gussets
(80)

Cruciform Joints (1)

Non load-carrying type
1. Fillet welded joints with smooth weld toes D(100)
2. Fillet welded joints with finished weld toes D(100)
3. As-welded fillet welded joints E(80)
4. Fillet welded joints including start and stop position E(80)

5. Fillet welded joints of F(65)

hollow section
G(50)

22
 Cruciform Joints (2)
Load-carrying type
D(100)
6. Complete
D(100)
penetration
weld as-welded E(80)
F(65)
E(80)
Fillet or partial penetration

7. Toe E(80)
failure as-welded F(65)
F(65)
8. Root failure H(40)

9. Hollow section H(40)

H(40)

Gusset Joints
1. Joints with fillet E(80)
1:(L≦100mm)
Out of plane gussets

welded or groove
welded gusset F(65) 3,4(L>100mm)
2. Joints with groove welded gusset with fillet E(80)

3. Joints with fillet welded gusset G(50)

4. Joints with F(65)

groove welded
gusset (L>100mm) G(50)
D(100)
In plane gussets

5. Joints with groove

welded gusset with E(80)
fillet
F(65)

6. Joints with groove G(50)

welded gusset
H(40)
7. Base plate with lap-welded gusset H(40)

23
Other Welded Joints

1. Joints with fillet E(80)

welded cover plates
(l<300mm) F(65)
2. Joints with fillet D(100)
welded cover plates
(l>300mm) G(50)
3. Welded studs E(80)
S(80)
H(40)

4. Lapped joints H(40)

H(40)
S(80)

Cables and High Strength Bolts

K1(270)
1. Cables
K2(200)
K1(270)
2. Cable
K2(200)
anchorages
K3(150)
3. High strength K4(65)
bolts K5(50)

24
 Correction Factor for Basic Allowable
Stress Ranges

Allowable Stress Range, Δσ R

Difference between real joints and the
experimental specimens in scale and residual
stress

Correction for allowable stress ranges

Allowable stress range = Basic allowable stress range
( Δσ R ) X CR X CT
Effect of
Effect of mean stress plate thickness

25
Effect of Mean Stress (1)
σ max
σ m (mean stress) = 1
2 (σ max + σ min )
σm σ min
σ min R (stress ratio) =
σ max

R >0 Variation of stress ratios, R

R =0
R = −1
R = −∞
Stress

Cycles

Max and min stress under D+L case are

used for calculation of R

Effect of Mean Stress (2)

Stress amplitude
R = −∞ R =0
R = −1
⎛ Δσ f ⎞
1.3 ⎜ ⎟
⎝ 2 ⎠ ⎛ Δσ f ⎞
⎜ 2 ⎟
⎝ ⎠

R>1 R≦-1 -1<R<1

0
Mean stress

⎛ 1−R ⎞
C R = 1.3 ⎜ ⎟ for R ≤ -1
⎝ 1.6 − R ⎠

26
Effect of Plate Thickness
Fatigue strength decreases with increase of
plate thickness in some kinds of joints
Example: Cruciform joints

C t = 4 25 for t > 25 mm
t
thickness

Stress Fluctuation
and
Stress Range Histograms

27
Stress Fluctuation
Strain responses due to running vehicle
running Stresses vary
At the bottom flange with positions of
of the main girder
loads

Variable
Stress records
amplitude
stresses

How to calculate
stress cycles ??

(4) Estimation of fatigue life,

Fatigue damage
rainflow method
stressσ ６０
stress range⊿σ

５０

４０

３０

２０
time
０ ５００ １０００

frequency

28
Rain Flow Counting Method
A method for determining a stress range histogram
from variable amplitude stresses
Origin point of rain drop
Analogy
The flow
of drops
of rain
down a
pagoda
roof.

Rainflow Method
cycles
2-3
5-6
4-7
8-9
11-12

open
1-10-13

29
Stress Range Histogram
Plot of Stress ranges and Frequencies obtained
from the rain flow counting method
Example: Oosaka Bridge, Japan

Strain gauge

Bottom flange
at mid span

Linear Damage Calculation

Cumulative Damage by 'Palmgren
'Palmgren--Miner"
Summation

Stress range histogram

S1 Cumulative damage, D
D = ∑ (ni N i )
stress

Si
n1 ni
frequency
Linear cumulative damage rule
proposed by Miner
S1 S-N curve

Si D=1 Failures
n1 N1 ni Ni

30
Equivalent Stress Range
Constant amplitude stress range, which causes fatigue damage
equivalent to the same repeated number of variable amplitude
stresses

Stress range histogram

Δσ e = m
∑ Δσ in
m
i i
S1
∑n i
stress

Δσ e
where
Si m = 3 for normal stress
n1 ni m = 5 for shear stress
frequency

Fatigue Assessment

31
Fatigue Design Load

T Load

Design specifications for highway bridge

Japan Road Association (JRA)

Safety Factors (γ b ⋅ γ w ⋅ γ i ) Δσ d ≤ Δ σ R

1. Redundancy factor, γb (0.8 - 1.1)

When damage occurs in the objective joint, it will
affect the whole structure strength

2. Importance factor, γw (0.8 - 1.1)

Degree of importance of a structure (social effect)

3. Inspection factor, γi (0.9 - 1.1)

Damage-detection probability by periodic inspections

Limitation

0.8 < γ b x γw x γ i < 1.25

32
 Fatigue Assessment
Based on Equivalent Stress Range
This equation should be satisfied.

(γ b ⋅ γ w ⋅ γ i ) Δσ d ≤ Δσ R
where
Δσ d = design stress range = equivalent stress range, Δσ e

Δσ R = allowable stress range

Basic of fatigue design in Steel highway bridge in

Japan

Fatigue design guideline for steel highway bridge

¾ Avoid low fatigue strength joints and joints

whose quality is uncontrolable.
¾ Use tough structural detail to fatigue : refer to
"Fatigue of steel bridge"
¾ Require quality control of welding to assure
fatigue strength. (Allowable defect size,
inspection area)
¾ Guideline is adopted to temporary members for
election, stiffening, etc..

33
 Flow of fatigue design
start
determine stress
range
• Avoid low fatigue strength
joints and joints whose
quality is uncontrollable. NG
• Use tough structural detail Δσ max ≤ Δσ CE ⋅ CR ⋅ Ct
to fatigue
nti = ADTTSLi ⋅ γ n ⋅ 365 ⋅ Y

OK Δσ i
−m
N i = C0 ⋅ ⎛⎜ ⎞
NO ⎝ C R Ct ⎟⎠
clear relation between nti
D=∑
acutual stress and i Ni
anlysed stress NG
OK
YES
detail design Δσ max maximum stress range change detail
end
Δσ CE cutoff limit for constant amp.
e.g. steel deck plate CR ⋅ Ct corection for average stress, thickness

Fatigue Design Curves from

IIW and AASHTO

34
IIW (International Institute of Welding : old)
Welded joints subjected Welded joints subjected
to Normal Stresses to Shear Stresses
14 categories 2 categories

IIW (International Institute of Welding : new)

Welded joints subjected to Normal Stress

14 categories

35
 American Association of State Highway and
AASHTO Transportation Officials

8 categories

m=3

Comparison of Strength Categories

Transverse butt welded joints
JSSC IIW ASSHTO

With ground flush surfaces 155 125 125

As-welded joint 100 100 89

Æ both side welds (toe angle = 30)

80
(other toe angle)

36
Comparison of Strength Categories
Longitudinal welded joints
JSSC IIW ASSHTO

Complete penetration 125 125 125

(without
groove welded joints from
stop/start
both sides positions)
Æ As-welded
90
(with
stop/start
positions)

Comparison of Strength Categories

Cruciform joints Æ Non load-carrying type
JSSC IIW ASSHTO

Fillet welded joints with 100 100 -

finished welded toes

As-welded fillet welded joints 80 80 89

37
Comparison of Strength Categories
Cruciform joints Æ Load-carrying type
JSSC IIW ASSHTO
Complete penetration weld 80 71 -
Æ as-welded

Fillet weld or Incomplete 65 63 -

penetration weld
Æ as-welded (Toe failure)

Comparison of Strength Categories

Gusset joints
Æ Out-of-plane gusset ASSHTO
JSSC IIW

Joints with fillet welded or 65 80 89

(L<50) (L<50)
groove welded gusset
(L<=100 mm)
Æ as-welded 71 71
(L<150) (50<L<100)

38
Comparison of Strength Categories
Gusset joints
Æ In-plane gusset ASSHTO
JSSC IIW

Joints with groove welded 40 50 89

(L<150) (L<50)
gusset
Æ as-welded
45 71
(L<300) (50<L<100)

40 For L>100
L (L>300) 56 (t<25)
40 (t>25)

Comparison of Strength Categories

Lapped joints
JSSC IIW ASSHTO

At base plates and splice 40 50 40

plates

39
assignment

40