IIWrec 2004 02

INTERNATIONAL INSTITUE INSTITUT INTERNATIONAL
OF WELDING DE LA SOUDURE
GERMAN DELEGATION DÉLÉGATION ALLEMANDE
Prof. Dr. A. Hobbacher

University of Applied Sciences, Friedrich-Paffrath-Str. 101, D-26389 Wilhelmshaven, Germany
Tel. +49 4421 985 518, Fax +49 4421 985 403, e-mail: [email protected]
IIW document XIII-1965-03 / XV-1127-03
RECOMMENDATIONS FOR FATIGUE DESIGN

OF WELDED JOINTS AND COMPONENTS
Update February 2004
This document was established by the IIW Joint Working Group XIII-XV
A. Hobbacher
Chairman of IIW JWG XIII-XV
IIW Fatigue Recommendations XIII-1965-02/XV-1127-03 February 2004
Hobbacher A.
Recommendations for Fatigue Design of Welded Joints and Components.
International Institute of Welding, doc. XIII-1965-03/XV-1127-03.
Paris, France, 2003
PREFACE
This document has been pepared as a result of an initiative by Commissions XIII and XV of the
International Institute of Welding (IIW). The task has been transferred to the Joint Working
Group XIII-XV, where it has been discussed and drafted in the years 1990 to 1996 and
updated in 2003. The document contains contributions from:
Prof. Dr. A. Hobbacher, Univ. Appl. Sc. Wilhelmshaven, Germany, as Chairman
Prof. Dr. H. Fricke, Univ. Hamburg-Harburg, Germany

Prof. P. Haagensen, Inst. of Technology, Trondheim, Norway
Prof. Dr. A. Hobbacher, Univ. of App. Sc., Wilhelmshaven, Germany
Mr. M. Huther, Bureau Veritas, France
Prof. Dr. K. Iida, Inst. of Technology, Shibaura, Japan
Dr. H.P. Lieurade, CETIM, Senlis, France
Dr. S.J. Maddox, The Welding Institute, Cambridge, U.K.
Prof. Dr. Ch. Miki, Inst. of Technology, Tokyo, Japan
Prof. Erkki Niemi, Lappeenranta Univ. of Technology, Finland
Mr. A. Ohta, NRIM, Tokyo, Japan
Mr. Oddvin Ørjasæter, SINTEF, Trondheim, Norway
Prof. Dr. H.J. Petershagen, Univ. Hamburg, Germany
Suggestions for a future refinement of the document are welcome and should be addressed to
the chairman:
Prof. Dr. A. Hobbacher

University of Applied Sciences
Friedrich-Paffrath-Str. 101
D-26389 Wilhelmshaven, Germany
e-mail: [email protected]
page 2
TABLE OF CONTENTS
1 GENERAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 SCOPE AND LIMITATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 DEFINITIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5 BASIC PRINCIPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 NECESSITY FOR FATIGUE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . 13
1.7 APPLICATION OF THE DOCUMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 FATIGUE ACTIONS (LOADING) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.1.1 Determination of Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.2 Stress Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.3 Types of Stress Raisers and Notch Effects . . . . . . . . . . . . . . . . . . . 18
2.2 DETERMINATION OF STRESSES + STRESS INTENSITY FACTORS . 20
2.2.1 Definition of Stress Components . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.2 Nominal Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.2.2 Calculation of Nominal Stress . . . . . . . . . . . . . . . . . . . . . 23
2.2.2.3 Measurement of Nominal Stress . . . . . . . . . . . . . . . . . . . 23
2.2.3 Structural Hot Spot Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.3.2 Types of hot spots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.3.3 Determination of Structural Hot Spot Stress . . . . . . . . . . 27
2.2.3.4 Calculation of Structural Hot Spot Stress . . . . . . . . . . . . 27
2.2.3.5 Measurement of Structural Hot Spot Stress . . . . . . . . . . . 31
2.2.3.6 Structural Hot Spot Stress Concentration Factors and
Parametric Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.4 Effective Notch Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.2.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.2.4.2 Calculation of Effective Notch Stress . . . . . . . . . . . . . . . 34
2.2.4.3 Measurement of Effective Notch Stress . . . . . . . . . . . . . . 35
2.2.5 Stress Intensity Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.2.5.2 Stress Intensity Factors by Parametric Formulae . . . . . . . 36
2.2.5.3 Stress Intensity Factors by Finite Elements . . . . . . . . . . . 37
2.3 STRESS HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3 FATIGUE RESISTANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2 FATIGUE RESISTANCE OF CLASSIFIED STRUCTURAL DETAILS . . 41
3.3 FATIGUE RESISTANCE AGAINST STRUCTURAL HOT SPOT
page 3
STRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.3.1 Fatigue Resistance using Reference S-N Curve . . . . . . . . . . . . . . . 74
3.3.2 Fatigue Resistance Using a Reference Detail . . . . . . . . . . . . . . . . . 75
3.4 FATIGUE RESISTANCE AGAINST EFFECTIVE NOTCH STRESS . . . . 77
3.4.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.5 FATIGUE STRENGTH MODIFICATIONS . . . . . . . . . . . . . . . . . . . . . . . . 78
3.5.1 Stress Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.5.1.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.5.1.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.5.2 Wall Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.5.2.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.5.2.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.5.3 Improvement Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.5.3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.5.3.2 Applicabiliy of Improvement Methods . . . . . . . . . . . . . . . 81
3.5.3.3 Burr Grinding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.5.3.4 TIG Dressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.5.3.5 Hammer Peening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.5.3.6 Needle Peening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.5.4 Effect of Elevated Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.5.4.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.5.4.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.5.5 Effect of Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.6 FATIGUE RESISTANCE AGAINST CRACK PROPAGATION . . . . . . . . 89
3.6.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.6.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.7 FATIGUE RESISTANCE DETERMINATION BY TESTING . . . . . . . . . . 90
3.7.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.7.2 Evaluation of Test Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.7.3 Evaluation of Data Collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3.8 FATIGUE RESISTANCE OF JOINTS WITH WELD IMPERFECTIONS . 94
3.8.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.8.1.1 Types of Imperfections . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.8.1.2 Effects and Assessment of Imperfections . . . . . . . . . . . . . 94
3.8.2 Misalignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.8.3 Undercut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.8.3.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.8.3.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.8.4 Porosity and Inclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.8.4.1 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.8.4.2 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.8.5 Cracklike Imperfections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.8.5.1 General Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
page 4
3.8.5.2 Simplified Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4 FATIGUE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

4.1 GENERAL PRINCIPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2 COMBINATION OF NORMAL AND SHEAR STRESS . . . . . . . . . . . . . 106
4.3 FATIGUE ASSESSMENT USING S-N CURVES . . . . . . . . . . . . . . . . . . 107
4.3.1 Linear Damage Calculation by "Palmgren-Miner" Summation . . . 107
4.3.2 Nonlinear Damage Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4 FATIGUE ASSESSMENT BY CRACK PROPAGATION CALCULATION112
4.5 FATIGUE ASSESSMENT BY SERVICE TESTING . . . . . . . . . . . . . . . . 113
4.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.5.2 Acceptance Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.5.3 Safe Life Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.5.4 Fail Safe Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4.5.5 Damage Tolerant Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5 SAFETY CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

5.1 BASIC PRINCIPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2 FATIGUE DESIGN STRATEGIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2.1 Infinite Life Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2.2 Safe Life Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.2.3 Fail Safe Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.2.4 Damage Tolerant Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.3 PARTIAL SAFETY FACTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.4 QUALITY ASSURANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.5 REPAIR OF COMPONENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6 APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1 LOAD CYCLE COUNTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.1 Transition Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.2 Rainflow or Reservoir Counting Method . . . . . . . . . . . . . . . . . . . 121
6.2 FRACTURE MECHANICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.2.1 Rapid Calculation of Stress Intensity Factors . . . . . . . . . . . . . . . . 122
6.2.2 Dimensions of Cracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.2.3 Interaction of Cracks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
6.2.4 Formulae for Stress Intensity Factors . . . . . . . . . . . . . . . . . . . . . . 124
6.3 FORMULAE FOR MISALIGNMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.4 STATISTICAL CONSIDERATIONS ON SAFETY . . . . . . . . . . . . . . . . . 135
6.4.1 Statistical Evaluation of Fatigue Test Data . . . . . . . . . . . . . . . . . . 135
6.4.2 Statistical Evaluation at Component Testing . . . . . . . . . . . . . . . . 136
6.4.3 Statistical Considerations for Partial Safety Factors . . . . . . . . . . . 138
7 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
page 5
1 GENERAL
The IIW, every other body or person involved in the preparation and publication of this
document hereby expressly disclaim any liability or responsibility for loss or damage resulting
from its use, for any violation of any mandatory regulation with which the document may
conflict, or for the infringement of any patent resulting from the use of this document.
It is the user's responsibility to ensure that the recommendations given here are suitable
for his purposes.
1.1 INTRODUCTION
The aim of these recommendations is to provide a basis for the design and analysis of welded
components loaded by fluctuating forces, to avoid failure by fatigue. In addition they may assist
other bodies who are establishing fatigue design codes. It is assumed that the user has a
working knowlegde of the basics of fatigue and fracture mechanics.
The purpose of designing a structure against the limit state due to fatigue damage is to ensure,
with an adequate survival probability, that the performance is satisfactory during the design life.
The required survival probability is obtained by the use of appropriate partial safety factors.
1.2 SCOPE AND LIMITATIONS

The recommendations present general methods for the assessment of fatigue damage in welded
components, which may affect the limit states of a structure, such as ultimate limit state and
servicability limited state [1].
The recommendations give fatigue resistance data for welded components made of wrought or
extruded products of ferritic/pearlitic or bainitic structural steels up to fy=960 MPa, of
austenitic stainless steels and of aluminium alloys commonly used for welded structures.
The recommendations are not applicable to low cycle fatigue, where )Fnom>1.5Afy ,
maxFnom>fy , for corrosive conditions or for elevated temperature operation in the creep
range.
page 6
1.3 DEFINITIONS
Characteristic value Loads, forces or stresses, which vary statistically, at a specified
fractile, here: 95% at a confidence level of the mean of 75% .
Classified structural
detail A structural detail containing a structural discontinuity
including a weld or welds, for which the nominal stress
approach is applicable, and which appear in the tables of the
recommendation. Also referred to as standard structural detail.
Concentrated load
effect A local stress field in the vicinity of a point load or reaction
force, or membrane and shell bending stresses due to loads
causing distortion of a cross section not sufficiently stiffened by
a diaphragm.
Constant amplitude
loading A type of loading causing a regular stress fluctuation with con-
stant magnitudes of stress maxima and minima.
Crack propagation
rate Amount of crack tip propagation during one stress cycle.
Crack propagation
threshold Limiting value of stress intensity factor range below which crack
propagation will not occur.
Cut off limit Fatigue strength under variable amplitude loading, below which
the stress cycles are considered to be non-damaging.
Design value Characteristic value factored by a partial safety factor.
Effective notch
stress Notch stress calculated for a notch with a certain effective notch
radius.
Equivalent stress
range Constant amplitude stress range which is equivalent in terms of
fatigue damage to the variable amplitude loading under study, at
the same number of cycles.
Fatigue Detoriation of a component caused by crack initiation and/or by

the growth of cracks.
page 7
Fatigue action Load effect causing fatigue.
Fatigue damage ratio Ratio of fatigue damage sustained to fatigue damage required to
cause failure, defined as the ratio of the number of applied stress
cycles and the corresponding fatigue life at constant amplitude.
Fatigue life Number of stress cycles of a particular magnitude required to

cause fatigue failure of the component.
Fatigue limit Fatigue strength under constant amplitude loading corre-

sponding to infinite fatigue life or a number of cycles large
enough to be considered infinite by a design code.
Fatigue resistance Structural detail's resistance against fatigue actions in terms of

S-N curve or crack propagation properties.
Fatigue strength Magnitude of stress range leading to a particular fatigue life.
Fracture mechanics A branch of mechanics dealing with the behaviour and strength
of components containing cracks.
Hot spot A point in a structure where a fatigue crack may initiate due to
the combined effect of structural stress fluctuation and the weld
geometry or a similar notch.
Local nominal stress Nominal stress including macro-geometric effects, concentrated

load effects and misalignments, disregarding the stress raising
effects of the welded joint itself. Also referred to as modified
nominal stress.
Local notch A notch such as the local geometry of the weld toe, including
the toe radius and the angle between the base plate surface and
weld reinforcement. The local notch does not alter the structural
stress but generates nonlinear stress peaks.
Macro-geometric
discontinuity A global discontinuity, the effect of which is usually not taken
into account in the collection of standard structural details, such
as a large opening, a curved part in a beam, a bend in a flange
not supported by diaphragms or stiffeners, discontinuities in
pressure containing shells, eccentricity in a lap joint (see fig.
(2.2)-3).
page 8
Macro-geometric effect A stress raising effect due to macro-geometry in the vicinity of

the welded joint, but not due to the welded joint itself.
Membrane stress Average normal stress across the thickness of a plate or shell.
Miner sum Summation of individual fatigue damage ratios caused by each

stress cycle or stress range block above a certain cut-off limit
according to the Palmgren-Miner rule.
Misalignment Axial and angular misalignments caused either by detail design

or by poor fabrication or welding distortion.
Modified nominal stress See 'Local nominal stress'.
Nominal stress A stress in a component, resolved using general theories, e.g.

beam theory. See also local nominal stress.
Nonlinear stress peak The stress component of a notch stress which exceeds the
linearly distributed structural stress at a local notch.
Notch stress Total stress at the root of a notch taking into account the stress
concentration caused by the local notch, consisting of the sum
of structural stress and nonlinear stress peak.
Notch stress concentration

factor The ratio of notch stress to structural stress.
Paris' law An experimentally determined relation between crack growth

rate and stress intensity factor range.
Palmgren-Miner rule
Fatigue failure is expected when the Miner sum reaches unity.
Rainflow counting A standardized procedure for stress range counting.
Range counting A procedure of determining various stress cycles and their

ranges from a stress history, preferably by rainflow counting
method.
Shell bending stress Bending stress in a shell or plate-like part of a component, li-
nearly distributed across the thickness as assumed in the theory
of shells.
S-N curve Graphical presentation of the dependence of fatigue life N on
page 9
fatigue strength S ()FR or )JR), also known as Wöhler curve.
Stress cycle A part of a stress history containing a stress maximum and a

stress minimum, determined usually by a range counting
method.
Stress history A time based presentation of a fluctuating stress, defined by

sequential stress peaks and troughs (valleys), either for the total
life or for a certain sample.
Stress intensity
factor Main parameter in fracture mechanics, the combined effect of
stress and crack size at the crack tip region.
Stress range The difference between stress maximum and stress minimum in
a stress cycle, the most important parameter governing fatigue
life.
Stress range block A part of the total spectrum of stress ranges which is discretized
in a certain number of blocks.
Stress range exceedances A tabular or graphical presentation of the cumulative frequency

of stress range exceedances, i.e the number of ranges exceeding
a particular magnitude of stress range in a stress history. Here,
frequency is the number of occurrances. (Also referred to as
"stress spectrum" or "cumulative frequency diagram").
Stress range occurrences A tabular or graphical presentation of stress ranges, usually

discretized in stress range blocks. See also "stress range excee-
dances".
Stress ratio Ratio of minimum to maximum algebraic value of the stress in

a particular stress cycle.
Stress intensity factor ratio Ratio of minimum to maximum algebraic value of the stress
intensity factor of a particular load cycle.
Structural discontinuity A geometric discontinuity due to the type of welded joint,

usually to be found in the tables of classified structural details.
The effects of a structural discontinuity are (i) concentration of
the membrane stress and (ii) formation of secondary shell
bending stresses (see fig. (2.2)-6).
Structural stress A stress in a component, resolved taking into account the

effects of a structural discontinuity, and consisting of membrane
page 10
and shell bending stress components. Also referred to as

geometric stress.
Structural stress
concentration factor The ratio of structural (hot spot) stress to modified (local)
nominal stress.
Structural hot spot stress The value of structural stress on the surface at a hot spot.
Variable amplitude loading A type of loading causing irregular stress fluctuation with stress
ranges (and amplitudes) of variable magnitude.
page 11
1.4 SYMBOLS
K stress intensity factor
Kmax stress intensity factor caused by Fmax
Kmin stress intensity factor caused by Fmin
Mk magnification function for K due to nonlinear stress peak
Mk,m magnification function for K, concerning membrane stresses
Mk,b magnification function for K, concerning shell bending stresses
R stress ratio
Y correction function for K, taking into account crack form, aspect ratio, relative
crack size etc.
Ym correction function for K, concerning membrane stress
Yb correction function for K, concerning shell bending stress
a depth of a surface crack or semi length of a through crack
ao initial depth of a surface crack
af crack size at failure
e eccentricity, amount of offset misalignment
fy actual or specified yield strength of the material
km stress magnification factor due to misalignment
ks stress concentration factor due to structural discontinuity
kt stress concentration factor due to local notch
m exponent of S-N curve or Paris power law
t plate thickness, thickness parameter (crack center to nearest surface)
)K stress intensity factor range
)KS,d design value of stress intensity factor range caused by actions
)Kth threshold stress intensity factor range
)F stress range
)FS,d design value of stress range caused by actions
)FR,L characteristic value of fatigue limit
)J shear stress range
(M partial safety factor for fatigue resistance in terms of stress
'M partial safety factor for fatigue resistance in terms of cycles
F normal stress
Fben shell bending stress
Fen effective notch stress Subscripts:
Fln (local) notch stress
Fmax stress maximum in stress history S fatigue actions
Fmem membrane stress R fatigue resistance
Fmin stress minimum in stress history
Fnlp nonlinear stress peak d design value
Fnom nominal stress k characteristic value
Fhs structural hot spot stress J shear stress
page 12
1.5 BASIC PRINCIPLES

According to the ISO format for verification of structures [1], fatigue action and fatigue
resistance are clearly separated. Fatigue resistance is given in terms of tentative data. The
representation of tentative data has also been separated from the assessment curves used for
damage calculation, because different damage calculation methods may require special
modifications to the resistance S-N curve, which is usually based on constant amplitude tests.
Thus, the flexibility and possibility for continuous updating of the document is maintained.
No recommendations are given for the fatigue load (action) side, nor for the partial safety
factor on fatigue actions (F.
The different approaches for the fatigue assessment of welded joints and components
considered are: nominal stress, Structural Stress, effective notch stress, fracture mechanics
method and component testing.
1.6 NECESSITY FOR FATIGUE ASSESSMENT

Fatigue assessment is generally required for components subject to fluctuating loads.
In the following cases, detailed fatigue assessment is not required:
a) The highest nominal design stress range satisfies
(M should be taken from an applicable design code.
This paragraph is not applicable to tubular joints.
b) A Miner sum (4.3.1) equal to D=0.5 using a FAT fatigue class according to
(3.2) of FAT 36 for steel or FAT 14 for aluminium corresponds to a fatigue life
greater than 5 million cycles.
c) For a detail for which a constant amplitude fatigue limit )FR,L is specified and
all design stress ranges are under the design resistance fatigue limit
d) For a crack, at which all design stress intensity factors are under the threshold
level )Kth for crack propagation.
page 13
for steel )Kth=2.0 MPa/m

for aluminium )Kth=0.7 MPa/m
1.7 APPLICATION OF THE DOCUMENT

Based on the initial information about the welded joint, the stress type to be used in the fatigue
assessment is defined and determined. The fatigue resistance data are then selected according
to the type of the fatigue action. The corresponding types of fatigue action and resistance are:
Tab. {1}-1: Presentation of fatigue actions and resistances vs. assessment procedure
Fatigue action Fatigue resistance Assessment procedure
Forces on Resistance determined by test at Component testing
component component
Nominal stress in Resistance given by tables of structural Summation of
section details in terms of a set of S-N curves cumulative damage
Structural hot spot Resistance against structural hot spot
stress at weld toe stress in terms of S-N curves
Effective notch Resistance against effective notch
stress in weld notch stress in terms of a universal S-N curve
Stress intensity at Resistance against crack propagation in Summation of crack
crack tip terms of the material parameters of the propagation
crack propagation law
The fatigue assessment procedure depends on the presentation of fatigue resistance data. The
chosen procedure has to be performed using adequate safety factors.
page 14
Tab. {1}-3: General guidance for the application of the document

Item Initial Infor- Fatigue Fatigue
mation Action Resistance
(1) Does joint determine look up go to (6)

correpond to a nominal stress fatigue
tabulated yes 6 (2.2.2) then 6 resistance
structural class (FAT)
detail? in tables
(3.2)
if no 9
(2) Is Structural determine look up re-

Stress assess- Structural sistance S-N
ment ap- yes 6 Stress (2.2.3) then 6 curve for go to (6)
plicable? Structural
Stress (3.3)
if no 9
(3) Is effective determine look up re-

notch stress effective sistance S-N
assessment yes 6 notch stress then 6 curve for go to (6)
applicable? (2.2.4) effective
notch stress
(3.4)
if no 9
(4) Are cracks or determine look up

cracklike stress inten- resistance
imperfections yes 6 sity factor then 6 against go to (7)
present? (2.2.5) crack pro-
pagation
(3.6 and 3.8)
if no 9
page 15
(5) Test entire go to (8)

component
(4.5)
test structural go to (1)
detail (3.7)
Modifications and Assessment Procedures

(6) Modify resis- is Miner calculate perform
tance S-N rule ad- design resis- summation
curve (3.5) for equate tance S-N then 6 (4.3.1)
all effects not (4.3)? curve (4.3.1) giving life
yet covered yes 6 using (M (8) cycles,
assess if OK
if no 6 calc. dimen-
sionless crack
propagation
param. from
resistance
S-N curve
(4.3.2) using
(M (8)
then 9
(7) calc. design perform crack

crack propa- propagation
gation resis- then 6 calc. (4.4)
tance data giving life
using 'M (8) cycles
assess if OK
Safety Considerations
(8) define (M according to safety considerations (chapter 5)
page 16
2 FATIGUE ACTIONS (LOADING)

All types of fluctuating load acting on the component and the resulting stresses at potential
sites for fatigue have to be considered. Stresses or stress intensity factors then have to be deter-
mined according to the fatigue assessment procedure applied.
The actions originate from live loads, dead weights, snow, wind, waves, pressure, ac-
celerations, dynamic response etc. Actions due to transient temperature changes should be
considered. Improper knowledge of fatigue actions is one of the major sources of fatigue
damage.
Tensile residual stresses due to welding decrease the fatigue resistance, however, the influence
of residual weld stresses is already included in the fatigue resistance data given in chapter 3.

2.1.1 Determination of Actions
The actions in service have to be determined in terms of characteristic loads. Partial safety
factors on actions (F have to be applied as specified in the application code giving the design
values of the actions for fatigue assessment.
In this document, there is no guidance given for the establishing of design values for actions
(loads), nor for partial safety factors (F for actions (loads).
2.1.2 Stress Range
Fatigue assessment is usually based on stress range or stress intensity factor range. Thus, the
actions have to be given in these terms.
page 17
The maximum and the minimum values of the stresses are to be calculated from a superposition
of all non permanent, i.e. fluctuating, actions:
a) fluctuations in the magnitudes of loads

b) movement of loads on the structure
c) changes in loading directions
d) structural vibrations due to loads and dynamic response
e) temperature transients
Fatigue analysis is based on the cumulative effect of all stress range occurrences during the
anticipated service life of the structure.
2.1.3 Types of Stress Raisers and Notch Effects
Different types of stress raisers and notch effects lead to the calculation of different types of
stress. The choice of stress depends on the fatigue assessment procedure used.
Tab. {2}-1: Stress raisers and notch effects
Type Stress raisers Stress determined Assessment procedure
A General analysis of Gross average not applicable for

sectional forces using stress from fatigue analysis, only
general theories e.g. beam sectional forces component testing
theory, no stress risers
considered
B A + macrogeometrical Range of nominal Nominal stress
effects due to the design of stress (also modi- approach
the component, but fied or local nomi-
excluding stress risers due nal stress)
to the welded joint itself.
C A + B + structural Range of Structural Stress (hot
discontinuities due to the structural spot stress) approach
structural detail of the Structural Stress
welded joint, but excluding (hot spot stress)
the notch effect of the weld
toe transition
D A + B + C + notch stress Range of elastic a) Fracture mechanics
concentration due to the notch stress (total approach
weld bead notches stress) b) effective notch
a) actual notch stress stress approach
b) effective notch stress
page 18
Fig. (2.1)-2 Notch stress and

Fig. (2.1)-1 Modified or local nominal structural stress
stress
page 19
2.2 DETERMINATION OF STRESSES AND STRESS

INTENSITY FACTORS
2.2.1 Definition of Stress Components
The stress distribution over the plate thickness is non-linear in the vicinity of notches.
Fig. (2.2)-1 Non-linear stress distribution separated to stress components
The stress components of the notch stress Fln are [2]:
Fmem membrane stress

Fnlp non-linear stress peak
If a refined stress analysis method is used, which gives a non-linear stress distribution, the stress
components can be separated by the following method:
The membrane stress Fmem is equal to the average stress calculated through the
thickness of the plate. It is constant through the thickness.
The shell bending stress Fben is linearly distributed through the thickness of the plate. It
is found by drawing a straight line through the point O where the membrane stress
intersects the mid-plane of the plate. The gradient of the shell bending stress is chosen
such that the remaining non-linearly distributed component is in equilibrium.
The non-linear stress peak Fnlp is the remaining component of the stress.
The stress components can be separated analytically for a given stress distribution F(x) for x=0
at surface to x=t at through thickness:
page 20
Fig. (2.2)-2 Position of coordinates
2.2.2 Nominal Stress
2.2.2.1 General
Nominal stress is the stress calculated in the sectional area under consideration, disregarding
the local stress raising effects of the welded joint, but including the stress raising effects of the
macrogeometric shape of the component in the vicinity of the joint, such as e.g. large cutouts.
Overall elastic behaviour is assumed.
The nominal stress may vary over the section under consideration. E.g. at a beam-like
component, the modified (also local) nominal stress and the variation over the section can be
calculated using simple beam theory. Here, the effect of a welded on attachment is ignored.
Fig. (2.2)-2 Nominal stress in a beam-like component
The effects of macrogeometric features of the component as well as stress fields in the vicinity
of concentrated loads must be included in the nominal stress. Consequently, macrogeometric
effects may cause a significant redistribution of the membrane stresses across the section.
Similar effects occur in the vicinity of concentrated loads or reaction forces. Significant shell
bending stress may also be generated, as in curling of a flange, or distortion of a box section.
page 21
Fig. (2.2)-3 Examples of macrogeometric effects
Fig. (2.2)-4 Modified (local) nominal stress near concentrated loads
The secondary bending stress caused by axial or angular misalignment needs to be considered
if the misalignment exceeds the amount which is already covered by fatigue resistance S-N
curves for the structural detail. This is done by the application of an additional stress raising
factor km,eff (see 3.8.2). Intentional misalignment (e.g.allowable misalignment specified in the
design stage) is considered when assessing the fatigue actions (stress) by multiplying by the
factor. If it is non-intentional, it is regarded as a weld imperfection which affects the fatigue
resistance and has to be considered by dividing the fatigue resistance (stress) by the factor.
page 22
Fig. (2.2)-5 Axial and angular misalignement
2.2.2.2 Calculation of Nominal Stress
In simple components the nominal stress can be determined using elementary theories of
structural mechanics based on linear-elastic behaviour. In other cases, finite element method
(FEM) modelling may be used. This is primarily the case in:
a) complicated statically over-determined (hyperstatic) structures

b) structural components incorporating macrogeometric discontinuities, for which
no analytical solutions are available
Using FEM, meshing can be simple and coarse. Care must be taken to ensure that all stress
raising effects of the structural detail of the welded joint are excluded when calculating the
modified (local) nominal stress.
If nominal stresses are calculated in fillet welds by a coarse finite element mesh, nodal forces
should be used in a section through the weld instead of element stresses in order to avoid
stress underestimation.
2.2.2.3 Measurement of Nominal Stress
The fatigue resistance S-N curves of classified structural details are based on nominal stress,
disregarding the stress concentrations due to the welded joint. Therefore the measured nominal
stress must exclude the stress or strain concentration due to the corresponding discontinuity in
the structural component. Thus, strain gauges must be placed outside of the stress
concentration field of the welded joint.
page 23
In practice, it may be necessary firstly to evaluate the extension and the stress gradient of the
field of stress concentration (see 2.2.3.4) due to the welded joint. For further measurements,
simple strain gauge application outside this field is sufficient.
2.2.3 Structural Hot Spot Stress
2.2.3.1 General
The structural or geometric stress Fhs at the hot spot includes all stress raising effects of a
structural detail excluding all stress concentrations due to the local weld profile itself. So, the
non-linear peak stress Fnlp caused by the local notch, i.e. the weld toe, is excluded from the
structural stress. The structural stress is dependent on the global dimensional and loading para-
meters of the component in the vicinity of the joint (type C in 2.1.3 table {2}-1). It is
determined on the surface at the hot spot of the component which is to be assessed. Structural
hot spot stresses Fhs are generally defined at plate, shell and tubular structures. Figure (2.2)-6
shows examples of structural discontinuities and details together with the structural stress
distribution.
Fig. (2.2)-6 Structural details and structural stress
The structural hot spot stress approach is recommended for welded joints where there is no
clearly defined nominal stress due to complicated geometric effects, and where the structural
page 24
discontinuity is not comparable to a classified structural detail.
Fig. (2.2)-7 Definition of structural hot spot stress
The structural hot spot stress can be determined using reference points and extrapolation to the
weld toe at the hot spot in consideration. The method as defined here is limited to the
assessment of the weld toe, i.e. cases a to e in fig.(2.2)-8. It is not applicable in cases where
crack will grow from the weld root and propagate through the weld metal, i.e. cases f to i in
fig. (2.2)-8.
Fig. (2.2)-8: Various locations of crack propagation in welded joints.
Note:
The method of structural hot spot stress may be extended to the assessment of spots of the
welded joint suceptible to fatigue cracking other than on plate surface, e.g. on a fillet weld
root. In this case, structural hot spot stress on surface is used as an indication and estimation
of the stress for the spot in consideration. The S-N curves or structural hot spot stress
concentration factors used for verification in this case depend largely on geometric and
page 25
Fig. (2.2)-9 Biaxial stress at weld toe

dimensional parameters and are only valid within the range of these parameters.
In case of a biaxial stress state at the plate surface, it is recommeded to use the principal stress
which is approximately in line with the perpendicular to the weld toe, i.e. within a deviation of
±60/ (fig. 2.2-9). The other principal stress may be analysed, if necessary, using the fatigue
class for parallel welds in the nominal stress approach.
2.2.3.2 Types of hot spots
Besides the definitions of structural hot spot stress as given above, two types of hot spots have
to be distiguished according to their location on the plate and their orientation to the weld toe:
Tab. {2.2}-1: Types of hot spots

Type Description Determination
a Structural hot spot stress transverse to Special FEA procedure or
weld toe on plate surface measurement and extrapolation
b Structural hot spot stress transverse to Special FEA procedure or
weld toe at plate edge measurement and extrapolation
page 26
2.2.3.3 Determination of Structural Hot Spot Stress
Determination of structural hot spot stress can be done either by measurement or by

calculation. Here the non-linear peak stress is eleminated by linearisation of the stress through
th plate thickness (see 2.2.1) or by extrapolation of the stress at the surface to the weld toe.
The following considerations focus on extrapolation procedures of the surface stress, which are
nearly the same in measurement and calculation.
Firstly the stresses at the reference points, i.e. extrapolation points, have to be determined,
secondly the structural hot spot stress has to be determined by extrapolation to the weld toe.
The structural hot spot stress may be determined using two or three stress or strain values at
particular reference points apart from the weld toe in direction of stress. The closest position
to the weld toe must be chosen to avoid any influence of the notch due to the weld itself (which
leads to a non-linear stress peak). This is practically the case at a distance of 0.4 t (t = plate
thickness) from the weld toe. The structural hot spot stress at the weld toe is then obtained by
extrapolation.
Identification of the critical points (hot spots) can be made by:
a) measuring several different points

b) analysing the results of a prior FEM analysis
c) experience of existing components, which failed
2.2.3.4 Calculation of Structural Hot Spot Stress
In general, analysis of structural discontinuities and details to obtain the structural hot spot
stress is not possible using analytical methods. Parametric formulae are rarely available. Thus,
finite element (FEM) analysis is mostly applied.
Type
Fig. (2.2)-10: Types of Hot Spots

page 27
Usually, structural hot spot stress is calculated on the basis of an idealized, perfectly aligned
welded joint. Consequently, any possible misalignment has to be taken explicitely into
consideration by the FEA model or by an appropriate stress magnification factor km, see also
3.8.2. This applies particularly to butt welds, cruciform joints and one-sided transverse fillet
welds at free, unsupported plates.
The extent of the finite element model has to be chosen such that constraining boundary effects
of the structural detail analysed are comparables to the acutal structure.
Models with thin plate or shell elements or alternatively with solid elements may be used. It
should be noted that on the one hand the arrangement and the type of the elements have to
allow for steep stress gradients as well as for the formation of plate bending, and on the other
hand, only the linear stress distribution in the plate thickness direction needs to be evaluated
with respect to the definition of the structural hot spot stress. The stresses should be
determined at the specified reference points.
For FEM analysis, sufficient expertise of the analyst is required. Guidance is given in [2]. In the
following, only some rough recommendations are given:
In a plate or shell element model (Fig. (2.2)-11, left part), the elements have to be arranged
in the mid-plane of the structural components. 8-noded elements are recommended particularly
in case of steep stress gradients. In simplified models, the welds are not modelled, except for
cases where the results are affected by local bending, e. g. due to an offset between plates or
due to the small distance between adjacent welds. Here, the welds may be included by vertical
or inclined plate elements having appropriate stiffness or by introducing constraint equations
or rigid links to couple node displacements.
An alternative particularly for complex cases is recommended using prismatic solid elements
which have a displacement function allowing steep stress gradients as well as plate bending
with linear stress distribution in the plate thickness direction. This is offered, e. g., by
isoparametric 20-node elements with mid-side nodes at the edges, which allow only one
element to be arranged in the plate thickness direction due to the quadratic displacement
function and the linear stress distribution. At a reduced integration, the linear part of the
stresses can be directly evaluated. Modelling of welds is generally recommended as shown in
fig. (2.2)-11 (right part).
The element lengths are determined by the reference points for the subsequent extrapolation.
In order to avoid an influence of the stress singularity, the stress closest to the hot spot is
usually evaluated at the first or second nodal point. Therefore, the length of the element at the
hot spot has to correspond at least to its distance from the first reference point. Coarser meshes
are possible with higher-order elements and fixed lengths, as further explained below.
Appropriate element widths are important particularly in cases with steep stress gradients. The
width of the solid element or the two shell elements in front of the attachment should not
page 28
exceed the attachment width 'w', i. e. the attachment thickness plus two weld leg lengths. See
also figure (2.2)-10.
Fig. (2.2)-11: Typical meshes and stress evaluation paths for a welded detail
Usually, the structural hot spot stress components are evaluated on the plate surface or edge.
Typical extrapolation paths are shown by arrows in fig. (2.2)-11. If the weld is not modelled,
it is recommended to extrapolate the stress to the structural intersection point in order to avoid
stress underestimation due to the missing stiffness of the weld.
Type “a” hot spots:
The structural hot spot stress Fhs is determined using the reference points and extrapolation
equations as given below (see also fig. (2.2)-12).
1) Fine mesh with element length not more than 0.4 t at the hot spot: Evaluation of nodal
stresses at two reference points 0.4 t and 1.0 t, and linear extrapolation (eq. 1).
(1)
2) Fine mesh as defined above: Evaluation of nodal stresses at three reference points 0.4 t,
0.9 t and 1.4 t, and quadratic extrapolation (eq. 2). This method is recommended in
cases with pronounced non-linear structural stress increase to the hot spot.
(2)
3) Coarse mesh with higher-order elements having lengths equal to plate thickness at the
hot spot: Evaluation of stresses at mid-side points or surface centers respectively, i.e.
at two reference points 0.5 t and 1.5 t, and linear extrapolation (eq. 3).
(3)
page 29
Type “b” hot spots:
The stress distribution is not dependent of plate thickness. So, the reference points are given at
absolute distances from the weld toe, or from the weld end if the weld does not continue
around the end of the attached plate.
4) Fine mesh with element length of not more than 4 mm at the hot spot: Evaluation of
nodal stresses at three reference points 4 mm, 8 mm and 12 mm and quadratic
extrapolation (eq. 4).
(4)
5) Coarse mesh with higher-order elements having length of 10 mm at the hot spot:
Evaluation of stresses at the mid-side points of the first two elements and linear extra-
polation (eq. 5).
(5)
Fig. (2.2)-12: Reference points at different types of meshing
page 30
Tab. 2.2.-1: Recommended meshing and extrapolation

Type of model Relatively coase models Relatively fine models
and weld toe
Type a Type b Type a Type b
Element Shells txt 10 x 10 mm #0.4 t x t or # 4 x 4 mm
size max t x w/2*) #0.4 t x w/2
Solids txt 10 x 10 mm #0.4 t x t or # 4 x 4 mm
max t x w #0.4 t x w/2
Extrapo- Shells 0.5 t and 1.5 t 5 and 15 mm 0.4 t and 1.0 t 4, 8 and 12 mm
lation mid-side mid-side points nodal points nodal points
points points**)
Solids 0.5 and 1.5 t 5 and 15 mm 0.4 t and 1.0 t 4, 8 and 12 mm
surface center surface center nodal points nodal points
*)
w = longitudinal attachment thichness + 2 weld leg lenths
**)
surface center at transverse welds, if the weld below the plate is not modelled (see left
part of fig. 2.2-11)
2.2.3.5 Measurement of Structural Hot Spot Stress
The recommended placement and number of strain gauges is dependent of the presence of
higher shell bending stresses, the wall thickness and the type of strucutral stress.
Fig. (2.2)-13: Examples of strain gauges in plate structures
The center point of the first gauge should be placed at a distance of 0.4 t from the weld toe.
The gauge length should not exceed 0.2 t. If this is not possible due to a small plate thickness,
the leading edge of the gauge should be placed at a distance 0.3 t from the weld toe. The
following extrapolation procedure and number of gauges are recommended:
page 31
Type “a” hot spots:
a) Two gauges at reference points 0.4 t and 1.0 t and linear extrapolation (eq. 6).
(6)
b) Three gauges at reference points 0.4 t, 0.9 t and 1.4 t, and quadratic extrapolation in
cases of pronounced non-linear structural stress increase to the hot spot (eq. 7).
(7)
Often multi-grid strip gauges are used with fixed distances between the gauges. Then the
gauges may not be located as recommended above. Then it is recommended to use e.g. four
gauges and fit a curve through the results.
Type “b” hot spots:
Strain gauges are attached at the plate edge at 4, 8 and 12 mm distant from the weld toe. The
hot spot strain is determined by quadratic extrapolation to the weld toe (eq. 8):
(8)
Tubular joints:
For tubular joints, there exist recommendations which allow the use of linear extrapolation
using two strain gauges. Here, the measurement of simple uniaxial stress is sufficient. For
additional details see ref. [xx]
Determination of stress:
If the stress state is close to uniaxial, the structural hot spot stress is obtained approximately
from eqn. (9).
(9)
At biaxial stress states, the actual stress may be up to 10% higher than obtained from eqn. (3).
In this case, use of rosette strain gauges is recommended. If FEA results are available giving
the ratio between longitudinal and transverse strains εy/εx , the structural hot spot stress σhs can
then be resolved from eqn. (10), assuming that this principal stress is about perpenticular to the
weld toe.
(10)
page 32
Instead of absolute strains, strain ranges ∆ε = εmax − εmin are usually measured and substituted
in the above equations, producing the range of structural hot spot stress ∆σhs .
2.2.3.6 Structural Hot Spot Stress Concentration Factors and Parametric Formulae
For many joints between circular section tubes parametric formulae have been established for
the stress concentration factor khs in terms of structural structural stress at the critical points
(hot spots). Hence the structural hot spot stress Fhs becomes:
(11)
where Fnom is the nominal axial membrane stress in the braces, calculated by elementary stress
analysis.
page 33
2.2.4 Effective Notch Stress
2.2.4.1 General
Effective notch stress is the total stress at the root of a notch, obtained assuming linear-elastic
material behaviour. To take account of the statistical nature and scatter of weld shape
parameters, as well as of the non-linear material behaviour at the notch root, the real weld
contour is replaced by an effective one. For structural steels an effective notch root radius of
r = 1 mm has been verified to give consistent results. For fatigue assessment, the effective
notch stress is compared with a common fatigue resistance curve.
The method is restricted to welded joints which are expected to fail from the weld toe or weld
root. Other causes of fatigue failure, e.g. from surface roughness or embedded defects, are not
covered. Also it is also not applicable where considerable stress components parallel to the
weld or parallel to the root gap exist.
The method is well suited to the comparison of alternative weld geometries. Unless otherwise
specified, flank angles of 30° for butt welds and 45° for fillet welds are suggested.
In cases where a mean geometrical notch root radius can be defined, e.g. after certain post weld
improvement procedures, this geometrical radius plus 1 mm may be used in the effective notch
stress analysis.
The method is limited to thicknesses t >= 5 mm. For smaller wall thicknesses, the method has
not yet been verified.
Note:
The method does not consider benign effects of welding residual stresses, as widely observed
at fillet welds roots which are not stress relieved. The results may be conservative in
comparison with nominal stress fatigue data which consider these residual stresses by
definition.
2.2.4.2 Calculation of Effective Notch Stress
Effective notch stresses or stress concentration factors can be calculated by parametric

formulae, taken from diagrams or calculated from finite element or boundary element models.
The effective notch radius is introduced such that the tip of the radius touches the root of the
real notch, e.g. the end of an unwelded root gap.
page 34
Fig. (2.2)-12 Effective notch stress concentration factors
Possible misalignment has to be considered in the calculations.
2.2.4.3 Measurement of Effective Notch Stress
Because the effective notch radius is an idealization, the effective notch stress cannot be
measured directly in the welded component. In contrast, the simple definition of the effective
notch can be used for photo-elastic stress measurements in resin models.
page 35
2.2.5 Stress Intensity Factors
2.2.5.1 General
Fracture mechanics assumes the existence of an initial crack ai. It can be used to predict the
growth of the crack to a final size af. Since for welds in structural metals, crack initiation
occupies only a small portion of the life, this method is suitable for assessment of fatigue life,
inspection intervals, crack-like weld imperfections and the effect of variable amplitude loading.
The parameter which describes the fatigue action at a crack tip in terms of crack propagation
is the stress intensity factor (SIF) range )K.
Fracture mechanics calculations generally have to be based on total stress at the notch root, e.g.
at the weld toe. For a variety of welded structural details, correction functions for the local
notch effect and the nonlinear stress peak of the structural detail have been established. Using
these correction functions, fracture mechanics analysis can be based on Structural hot spot
stress or even on nominal stress. The correction function formulae may be based on different
stress types. The correction function and the stress type have to correspond.
2.2.5.2 Calculation of Stress Intensity Factors by Parametric Formulae
First, the local nominal stress or the structural Structural hot spot stress at the location of the
crack has to be determined, assuming that no crack is present. The stress should be separated
into membrane and shell bending stresses. The stress intensity factor (SIF) K results as a
superposition of the effects of both stress components. The effect of the remaining stress
raising discontinuity or notch (non-linear peak stress) has to be covered by additional factors
Mk.
where
K stress intensity factor
Fmem membrane stress
Ymem correction function for membrane stress intensity factor
Yben correction function for shell bending stress intensity factor
Mk, mem correction for non-linear stress peak in terms of membrane action
Mk, ben correction for non-linear stress peak in terms of shell bending
The correction functions Ymem and Yben can be found in the literature. The solutions in ref. [14-
16] are particularly recommended. For most cases, the formulae for stress intensity factors
given in appendix 6.2 are adequate. Mk-factors may be found in references [19] and [20].
page 36
2.2.5.3 Calculation of Stress Intensity Factors by Finite Elements
Stress intensity factor determination methods are usually based on FEM analyses. They may be
directly calculated as described in the literature, or indirectly using the weight function
approach. For more details see appendix 6.2
page 37
2.3 STRESS HISTORY

2.3.1 General
The fatigue design data presented in chapter 3 were obtained from tests performed under
constant amplitude loading. However, loads and the resulting fatigue actions (i.e. stresses) on
real structures usually fluctuate in an irregular manner and give rise to variable amplitude
loading. The stress amplitude may vary in both magnitude and periode from cycle to cycle.
Because the physical phemomena implied in fatigue are not chaotic but random, the
fluctuations can be characterized by mathematical func-tions and by a limited number of
parameters.
The stress history is a record and/or a representation of the fluctuations of the fatigue actions
in the anticipated service time of the component. It is described in terms of successive maxima
and minima of the stress caused by the fatigue actions. It covers all loading events and the
corresponding induced dynamic response.
Fig. (2.3)-1 Stress time history illustration
In most cases, the stress-time history is stationary and ergodic, which allows the definition of
a mean range and its variance, a statistical histogram and distribution, an energy spectrum and
a maximum values probabilistic distribution from a representation of a limited length.
Therefore, the data needed to perform a fatigue analysis can be determined from measurements
conducted during a limited time.
A stress history may be given as
a) a record of successive maxima and minima of stress measured in a comparable structure

with comparable loading and service life, or a typical sequence of load events.
b) a two dimensional transition matrix of the stress history derived from a).
c) a one- or two-dimensional stress range histogram (stress range occurrences) obtained

from a) by a specified counting method.
page 38
d) a one-dimensional stress range histogram (stress range exceedances, stress range

spectrum) specified by a design code.
The representations a) and b) may be used for component testing. c) and d) are most useful for
fatigue analysis by calculation.
2.3.2 Cycle Counting Methods

Cycle counting is the process of converting a variable amplitude stress sequence into equivalent
(in terms of fatigue damage) constant amplitude stress cycles. Different methods of counting
are in use, e.g. zero crossing counting, peak counting, range pair counting or rainflow
counting. For welded components, the reservoir or rainflow method is recommended for
counting stress ranges [26 and 27].
2.3.3 Cumulative Frequency Diagram (Stress Spectrum)
The cumulative frequency diagram (stress spectrum) corresponds to the cumulative probability
of stress range expressed in terms of stress range level exceedances versus the number of
cycles. The curve is therefore continuous.
The spectrum may be discretized giving a table of discrete blocks. For damage calculations 20
stress levels are recommended if more than 108 cycles are expected. Below this number of
cycles, 8 or 10 stress levels may be sufficient. All cycles in a block should be assumed to be
equal to the mean of the stress ranges in the block.
Besides the representation in probabilities, a presentation of the number of occurrences or

exceedances in a given number of cycles, e.g. 1 million, is used. An example showing a
Gaussian normal distribution is given below:
Tab. {2.3}-1: Stress range occurrance table (stress histogram or frequency)
# of block Relative Occurrence

stress range (frequency)
1 1.000 2
2 0.950 16
3 0.850 280
4 0.725 2720
5 0.575 20000
6 0.425 92000
7 0.275 280000
8 0.125 605000
page 39
Fig. (2.3)-2 Cumulative frequency diagram (stress spectrum)
page 40
3 FATIGUE RESISTANCE

Fatigue resistance is usually derived from constant or variable amplitude tests. The fatigue
resistance data given here are based on published results from constant amplitude tests.
Guidance on the direct use of test data is given in section 3.7 and 4.5.
The fatigue resistance data must be expressed in terms of the same stress as that controlled or
determined during the generation of those data.
In conventional endurance testing, there are different definitions of failure. In general, small
specimens are tested to complete rupture, while in large components the observation of a
through wall crack is taken as a failure criterion. The fatigue resistance data are based on the
number of cycles N to failure. The data are represented in S-N curves
In fracture mechanics crack propagation testing, the crack growth rate data are derived from
crack propagation monitoring.
All fatigue resistance data are given as characteristic values, which are assumed to have a
survival probability of at least 95%, calculated from a mean value of a two-sided 75%
confidence level, unless otherwise stated (see 3.7).
The (nominal) stress range should be within the limits of the elastic properties of the material.
stresses or 1.5 A fy/%3 for nominal shear stresses.

The range of the design values of the stress range shall not exceed 1.5 A fy for nominal normal
3.2 FATIGUE RESISTANCE OF CLASSIFIED STRUC-

TURAL DETAILS
The fatigue assessment of classified structural details and welded joints is based on the nominal
stress range. In most cases structural details are assessed on the basis of the maximum principal
stress range in the section where potential fatigue cracking is considered. However, guidance
is also given for the assessment of shear loaded details, based on the maximum shear stress
range. Separate S-N curves are provided for consideration of normal or shear stress ranges, as
illustrated in figures (3.2)-1 and (3.2)-2 respectively.
page 41
Care must be taken to ensure that the stress used for the fatigue assessment is the same as that
given in the tables of the classified structural details. Macro-structural hot spot stress con-
centrations not covered by the structural detail of the joint itself, e.g. large cutouts in the
vicinity of the joint, have to be accounted for by the use of a detailed stress analysis, e.g. finite
element analysis, or appropriate stress concentration factors (see 2.2.2).
The fatigue curves are based on representative experimental investigations and thus include the
effects of:
structural hot spot stress concentrations due to the detail shown

local stress concentrations due to the weld geometry
weld imperfections consistent with normal fabrication standards
stress direction
welding residual stresses
metallurgical conditions
welding process (fusion welding, unless otherwise stated)
inspection procedure (NDT), if specified
postweld treatment, if specified
Furthermore, within the limits imposed by static strength considerations, the fatigue curves of
welded joints are independent of the tensile strength of the material.
Each fatigue strength curve is identified by the characteristic fatigue strength of the detail at 2
million cycles. This value is the fatigue class (FAT).
The slope of the fatigue strength curves for details assessed on the basis of normal stresses (fig.
(3.2)-1) is m=3.00. The constant amplitude fatigue limit is 5A 106 cycles. The slope of the
fatigue strength curves for detailes assessed on the basis of shear stresses (fig. (3.2)-2) is
m=5.00, but in this case the fatigue limit corresponds to an endurance of 108 cycles.
page 42
Fig. (3.2)-1: Fatigue resistance S-N curves for m=3.00, normal stress (steel)
Fig. (3.2)-2 Fatigue resistance S-N curves for shear stress (steel)
page 43
The descriptions of the structural details only partially include information about the weld size,
shape and quality. The data refer to a standard quality as given in codes and standard welding
procedures. For higher or lower qualities, conditions of welding may be specified and veryfied
by test (3.7).
The fatigue classses given in table {3.2-1} shall be modified as given in 3.5. The limitations of
weld imperfections shall be considered (3.8).
All butt welds shall be full penetration welds without lack of fusion, unless otherwise stated.
All S-N curves of details are limited by the material S-N curve, which may vary due to different
strengths of the materials.
Disregarding major weld defects, fatigue cracks originate from the weld toe, and then propa-
gate through the base material, or from the weld root, and then propagate through the weld
throat. For potential toe cracks, the nominal stress in the base material has to be calculated and
compared with the fatigue resistance given in the tables. For potential root cracks, the nominal
stress in the weld throat has to be calculated. If both failure modes are possible, e.g. at cruci-
form joints with fillet welds, both potential failure modes have to be assessed.
page 44
Fig. (3.2)-3: Fatigue resistance S-N curves for m=3.00, normal Fig. (3.2)-4 Fatigue resistance curves for aluminium (normal stress)
stress (steel)
page 45
Tab. {3.2}-1: Fatigue resistance values for structural details in steel and aluminium assessed on the basis of nominal stresses.
No. Structural Detail Description FAT FAT Requirements and Remarks

(St.= steel; Al.= aluminium) St. Al.
100 Unwelded parts of a component
111 Rolled or extruded products, compo- No fatigue resistance of any detail to be higher at any
nents with mashined edges, seamless number of cycles!
hollow sections.
Sharp edges, surface and rolling flaws to be removed
m=5 by grinding. Any machining lines or groves to be par-
allel to stresses!
St.: For high strength steels a hig- 160
her FAT class may be used if For high strength steels a higher FAT class may be
verified by test. used if verified by test.
Al.: AA 5000/6000 alloys 70

AA 7000 alloys 80
121 Machine gas cut or sheared material 140 --- All visible signs of edge imperfections to be removed.
with subsequent dressing, no cracks by The cut surfeces to be mashined or ground, all burrs to
inspection, no visible imperfections be removed.
m=3 No repair by welding refill!
Notcheffects due to shape of edges have to be conside-

red.
page 46

122 Machine thermally cut edges, corners 125 40 Notcheffects due to shape of edges have to be conside-
removed, no cracks by inspection red.
m=3
123 Manually thermally cut edges, free 100 --- Notcheffects due to shape of edges have to be conside-
from cracks and severe notches red.
m=3
124 Manually thermally cut edges, uncon- 80 --- Notcheffects due to shape of edges have to be conside-
trolled, no notch deeper red.
than .5 mm
m=3
200 Butt welds, transverse loaded
page 47

211 Transverse loaded butt weld (X-groove 100 40 All welds ground flush to surface, grinding paralell to
or V-groove) ground flush to plate, direction of stress. Weld run-on and run-off pieces to
100% NDT be used and subsequently removed. Plate edges to be
Discussion: EC3 has 125 ?? ground flush in direction of stress. Welded from both
sides. No misalignement.
Required quality cannot be inspected by NDT !

212 Transverse butt weld made in shop in 90 36 Weld run-on and run-off pieces to be used and subse-
flat position, quently removed. Plate edges to be ground flush in
toe angle # 30°, NDT direction of stress.
Discussion: EC3 has 112??
Welded from both sides. Misalignment <5%
213 Transverse butt weld not satisfying 80 Weld run-on and run-off pieces to be used and subse-
conditions of 212, NDT quently removed. Plate edges to be ground flush in
direction of stress.
Al.: Butt weld with toe angle #50° 32
Butt welds with toe angle >50/ 25 Welded from both sides. Misalignment <10%
page 48

214 Transverse butt weld, welded on cer- 80 28- Backing removed, root visually inspected.
amic backing, root crack Misalignment <10%
215 Transverse butt weld on permanent 71 25 Misalignment <10%

backing bar terminating >10 mm from
plate edge, else 63 22
216 Transverse butt welds welded from one Misalignment <10%

side without backing bar, full penetrati-
on
root controlled by NDT 71 28

no NDT 36 12
page 49

217 Transverse partial penetration butt 36 12 The detail is not recommended for fatigue loaded
weld, analysis based on stress in weld members.
throat sectional area, weld overfill not
to be taken into account. Assessment by notch stress or fracture mechanics is
preferred.
221 Transverse butt weld ground flush, All welds ground flush to surface, grinding paralell to
NDT, with transition in thickness and direction of stress. Weld run-on and run-off pieces to
width be used and subsequently removed. Plate edges to be
slope 1:5 100 40 ground flush in direction of stress.
slope 1:3 90 32
slope 1:2 80 28 Misalignment <10%
Exceeding misalignment due to thickness step to be

considered, see chapter 3.8.2
222 Transverse butt weld made in shop, Weld run-on and run-off pieces to be used and subse-
welded in flat position, weld profile quently removed. Plate edges to be ground flush in
controlled, NDT, with transition in direction of stress.
thickness and width:
slope 1:3 80 28
slope 1:2 72 25 Exceeding misalignment due to thickness step to be
page 50

223 Transverse butt weld, NDT, with tran- Weld run-on and run-off pieces to be used and subse-
sition on thickness and width quently removed. Plate edges to be ground flush in
slope 1:5 80 25 direction of stress.
slope 1:3 71 22
Exceeding misalignments due to thickness step to be

224 Transverse butt weld, different thick- 71 22 Misalignment <10% of smaller plate thickness
nesses without transition,
centres aligned.
In cases, where weld profile is equiva-
lent to a moderate slope transition, see
no. 222
225 Three plate connection, root crack 71 22 Arc welds: Misalignment <10%
page 51

226 Transverse butt weld flange splice in 100 40 All welds ground flush to surface, grinding paralell to
built-up section welded prior to the as- direction of stress. Weld run-on and run-off pieces to
sembly, ground flush, with radius tran- be used and subsequently removed. Plate edges to be
sition, NDT ground flush in direction of stress.
231 Transverse butt weld splice in rolled 80 28 All welds ground flush to surface, grinding paralell to
section or bar besides flats, ground direction of stress. Weld run-on and run-off pieces to
flush, NDT be used and subsequently removed. Plate edges to be
ground flush in direction of stress.
232 Transverse butt weld splice in circular Welded in flat position.

hollow section, welded from one side,
full penetration, root crack
root inspected by NDT 71 28

no NDT 36 12
233 Tubular joint with permanent backing 71 28 Welded in flat position.
page 52

234 Transverse butt weld splice in rectan- Welded in flat position.

gular hollow section, welded from one
side, full penetration, root crack
root inspected by NDT 56 25

no NDT 36 12
241 Transverse butt weld ground flush, 100 40 All welds ground flush to surface, grinding paralell to
weld ends and radius ground, 100% direction of stress. Weld run-on and run-off pieces to
NDT at crossing flanges, radius be used and subsequently removed. Plate edges to be
transition. ground flush in direction of stress.
Welded from both sides. No misalignment. Required

weld quality cannot be inspected by NDT
242 Transverse butt weld made in shop at 90 36 Weld run-on and run-off pieces to be used and sub-
flat position, weld profile controlled, sequently removed. Plate edges to be ground flush in
NDT, at crossing flanges, radius transi- direction of stress.
tion
Welded from both sides.Misalignment <5%
page 53

243 Transverse butt weld ground flush, 80 32 All welds ground flush to surface, grinding paralell to
NDT, at crossing flanges with welded direction of stress. Plate edges to be ground flush in
triangular transition plates, weld ends direction of stress.
ground.
Crack starting at butt weld. Welded from both sides. Misalignment <10%
For crack of throughgoing flange see
details 525 and 526!
244 Transverse butt weld, NDT, at crossing 71 28 Plate edges to be ground flush in direction of stress.
flanges, with welded triangular transi-
tion plates, weld ends ground. Welded from both sides.Misalignment <10%
Crack starting at butt weld.

245 Transverse butt weld at crossing 50 20 Welded from both sides.Misalignment <10%
flanges.
Crack starting at butt weld.

246 picture needed! Laser beam butt weld
300 Longitudinal load-carrying welds
page 54

311 Automatic longitudinal seam welds 125 50

without stop/start positions in hollow
sections
with stop/start positions 90 36
312 Longitudinal butt weld, both sides 125 50 Discussion: What is the use of NDT??
ground flush parallel to load direction,
100% NDT
313 Longitudinal butt weld, without 125 50

stop/start positions, NDT
with stop/start positions 90 36
321 Continuous automatic longitudinal 125 50 No start-Stop position is permitted except when the
fully penetrated K-butt weld without repair is performed by a specialist and inspection is
stop/start positions (based on stress carried out to verify the proper execution of the weld.
range in flange) NDT
page 55

322 Continuous automatic longitudinal 100 40 Discussion: EC3 has 112 ??

double sided fillet weld without stop-
/start positions (based on stress range
in flange)
323 Continuous manual longitudinal fillet 90 36

or butt weld (based on stress range in
flange)
324 Intermittent longitudinal fillet weld Analysis based on normal stress in flange and shear
(based on normal stress in flange F and stress in web at weld ends.
shear stress in web J at weld ends).
J/ F = 0
0.0 - 0.2 80 32 representation by formula??
0.2 - 0.3 71 28
0.3 - 0.4 63 25 steel but >=36
0.4 - 0.5 56 22
0.5 - 0.6 50 20
0.6 - 0.7 45 18
alum. but >=14
> 0.7 40 16
36 14
page 56

325 Longitudinal butt weld, fillet weld or Analysis based on normal stress in flange and shear
intermittent weld with cope holes stress in web at weld ends.
(based on normal stress in flange F and
shear stress in web J at weld ends), representation by formula??
cope holes not higher than 40% of web.
J/ F = 0 steel but >=36
0.0 - 0.2 71 28
0.2 - 0.3 63 25
0.3 - 0.4 56 22
alum. but >=14
0.4 - 0.5 50 20
0.5 - 0.6 45 18
> 0.6 40 16
36 14
page 57

331 Joint at stiffened knuckle of a flange to --- ---

be assessed according to no. 411 - 414,
depending on type of joint.
Stress in stiffener plate:
Af = area of flange
ASt = area of stiffener
Stress in weld throat:
Aw = area of weld throat
page 58

332 Unstiffened curved flange to web joint, --- --- The resulting force of Ff-left and Ff-right will bend the
to be assessed according to no. 411 - flange perpenticular to the plane of main loading. In
414, depending on type of joint. oder to minimize this additional stressing of the welds,
it is recommended to minimize the width and to
Stress in web plate: maximize the thickness of the flange.
Stress longitudinally to the weld is to be considered. At

additional shear, principle stress in web is to be
consired (see 321 to 323)
Stress in weld throat:
Ff axial force in flange

t thickness of web plate
a weld throat
400 Cruciform joints and/or T-joints
411 Cruciform joint or T-joint, K-butt 80 28 Material quality of intermediate plate has to be
welds, full penetration, no lamellar checked against susceptibility of lamellar tearing.
tearing, misalignment e<0.15At, weld
toes ground, toe crack Misalignment <15% of primary plate.
page 59

412 Cruciform joint or T-joint, K-butt 71 25 Material quality of intermediate plate has to be
welds, full penetration, no lamellar checked against susceptibility of lamellar tearing.
tearing, misalignment e<0.15At, toe
crack Misalignment <15% of primary plate.
413 Cruciform joint or T-joint, fillet welds 63 22 Material quality of intermediate plate has to be
or partial penetration K-butt welds, no checked against susceptibility of lamellar tearing.
lamellar tearing, misalignment
e<0.15At, Misalignment <15% of primary plate.
toe crack
414 Cruciform joint or T-joint, fillet welds 36 12 Analysis based on stress in weld throat.
or partial penetration K-butt welds
including toe ground joints,
weld root crack.
For a/t<=0.3 40 14 Ratio a/t is calculated from weld throat over wall
thicknes Discussion!!
page 60

415 Cruciform joint or T-joint, single-

sided arc or laser beam welded
V-butt weld, full penetration, no
lamellar tearing, misalignment e<
0.15At, toe crack. Root inspected. 71 25
If root is not inspected, then root 36 12

crack
416 Cruciform joint or T-joint, single-sided 71 25 Analysis based on stress in weld root. Excentricity e of
arc welded fillet or partial penetration plate t and weld throat a midpoints to be considered in
Y-butt weld, no lamellar tearing, analysis. Stress at weld root:
misalignment of plates e< 0.15At, stress
at weld root. Penetration verified. )F w, root = )F w, nom A (1+6e/a)
Penetration not verified. 36 12 An analysis by effective notch procedure is

recommended
421 Splice of rolled section with inter- 36 12
mediate plate, fillet welds,
weld root crack.
Analysis base on stress in weld throat.
page 61

422 Splice of circular hollow section with Welds NDT inspected in order to ensure full root
intermediate plate, singlesided butt penetration.
weld, toe crack
wall thickness > 8 mm 56 22
wall thickness < 8 mm 50 20
423 Splice of circular hollow section with

intermediate plate, fillet weld, root
crack. Analysis based on stress in weld
throat.

424 Splice of rectangular hollow section, Welds NDT inspected in order to ensure full root
single-sided butt weld, toe crack penetration.
425 Splice of rectangular hollow section

with intermediate plate, fillet welds,
root crack
page 62

431 Weld connecting web and flange, --- --- Full penetration butt weld.
loaded by a concentrated force in web
plane perpendicular to weld. Force
distributed on width b = 2Ah + 50 mm.
Assessment according to no. 411 - 414.
A local bending due to eccentric load
should be considered.
500 Non-load-carrying attachments
511 Transverse non-load-carrying Grinding parallel to stress

attachment, not thicker than main plate
K-butt weld, toe ground 100 36
Two-sided fillets, toe ground 100 36
Fillet weld(s), as welded 80 28 At one sided fillet welds, an angular misalignment
thicker than main plate 71 25 corresponding to km = 1.2 is already covered
512 Transverse stiffener welded on girder For weld ends on web principle stress to be used
web or flange, not thicker than main
plate.
Two-sided fillets, toe ground 100 36
fillet weld(s): as welded 80 28
thicker than main plate 71 25
page 63

513 Non-loadcarrying stud 80 28

as welded
514 Trapezoidal stiffener to deck plate, full 71 25

penetration butt weld, calculated on
basis of stiffener thickness, out of plane
bending
515 Trapezoidal stiffener to deck plate, 50 16 Calculation on basis of stiffener thickness and weld
fillet or partial penetration weld, out of throat, whichever is smaller
plane bending
521 Longitudinal fillet welded gusset at For gusset near edge: see 525 "flat side gusset"
length l
l < 50 mm 80 28 If attachement thickness < 1/2 of base plat thickness,
l < 150 mm 71 25 then one step higher allowed (not for welded on
l < 300 mm 63 20 profiles!)
l > 300 mm 50 18
page 64

522 Longitudinal fillet welded gusset with 90 32 t = thickness of attachment

radius transition, end of fillet weld
reinforced and ground,
c < 2 t, max 25 mm
r > 150 mm
523 Longitudinal fillet welded gusset with t = thickness of attachment

smooth transition (sniped end or
radius) welded on beam flange or plate. If attachement thickness < 1/2 of base plat thickness,
c < 2 t, max 25 mm then one step higher allowed (not for welded on
profiles!)
r > 0.5 h 71 25
r < 0.5 h or n < 20° 63 20
524 Longitudinal flat side gusset welded on t = thickness of attachment
plate edge or beam flange edge, with
smooth transition (sniped end or For t2 < 0.7 t1, FAT rises 12%
radius). c < 2t2, max. 25 mm
r > 0.5 h 50 18
r < 0.5 h or n < 20° 45 16
page 65

525 Longitudinal flat side gusset welded on For t2 < 0.7 t1, FAT rises 12%
plate or beam flange edge, gusset
length l:
l < 150 mm 50 18
l < 300 mm 45 16
l > 300 mm 40 14
526 Longitudinal flat side gusset welded on Smooth transition radius formed by grinding the weld
edge of plate or beam flange, radius area in transition in order to temove the weld toe
transition ground. completely. Grinding parallel to stress.
r>150 or r/w > 1/3 90 36
1/6 < r/w < 1/3 71 28
r/w < 1/6 50 22
531 Circular or rectangular hollow section, 71 28 Non load-carrying welds. Width parallel to stress
fillet welded to another section. Section direction < 100 mm.
width parallel to stress direction < 100
mm, else like longitudinal attachment
600 Lap joints
611 Transverse loaded lap joint with fillet Stresses to be calculated in the main plate using a plate
welds width equalling the weld length.
Fatigue of parent metal 63 22 Buckling avoided by loading or design!
Fatigue of weld throat 45 16
page 66

612 Longitudinally loaded lap joint with Weld terminations more than 10 mm from plate edge.
side fillet welds Buckling avoided by loadin or design!
Fatigue of parent metal 50 18
Fatigue of weld (calc. on max.
weld length of 40 times the 50 18
throat of the weld
613 Lap joint gusset, fillet welded, non- t = thickness of gusset plate
load-carrying, with smooth transition
(sniped end with n<20° or radius),
welded to loaded element c<2At,
but c <= 25 mm
to flat bar 63 22
to bulb section 56 20
to angle section 50 18
614 Transverse loaded overlap joint with Stresses to be calculated using a plate width equalling
fillet welds. the weld length.
Fatigue of plate 63 22 For stress in plate, excenticity to be considered, as
Discussion: New detail!! given in chapters 3.8.2 and 6.3.
Fatigue of weld throat 36 12
700 Reinforcements
711 End of long doubling plate on I-beam, End zones of single or multiple welded cover plates,
welded ends (based on stress range in with or without frontal welds.
flange at weld toe)
tD # 0.8 t 56 20 If the cover plate is wider than the flange, a frontal
0.8 t < tD # 1.5 t 50 18 weld is needed. No undercut at frontal welds!
tD > 1.5 t 45 16
page 67

712 End of long doubling plate on beam, Grinding parallel to stress direction.
reinforced welded ends ground (based
on stress range in flange at weld toe)
tD # 0.8 t 71 28
0.8 t < tD # 1.5 t 63 25
tD > 1.5 t 56 22
721 End of reinforcement plate on 50 20 No undercut at frontal weld!

rectangular hollow section.
wall thickness:
t < 25 mm
731 Reinforcements welded on with fillet Grinding in direction of stress!

welds, toe ground 80 32
Toe as welded 71 25 Analysis based on modified nominal stress, however,
structural stress approach recommended.
800 Flanges, branches and nozzles
page 68

811 Stiff block flange, full penetration weld 71 25
812 Stiff block flange, partial penetration

or fillet weld
toe crack in plate 63 22
root crack in weld throat 36 12
821 Flat flange with > 80% full penetration 71 25

butt welds, modified nominal stress in
pipe, toe crack
822 Flat flange with fillet welds, modified 63 22

nominal stress in pipe, toe crack.
page 69

831 Tubular branch or pipe penetrating a 80 28 If diameter > 50 mm, stress concentration of cutout has
plate, K-butt welds. to be considered
Assessment by structural hot spot is recommended.
832 Tubular branch or pipe penetrating a 71 25 If diameter > 50 mm, stress concentration of cutout has
plate, fillet welds. Toe cracks. to be considered
Root cracks (analysis based on stress in 36 12 Assessment by structural hot spot is recommended.
weld throat)
841 Nozzle welded on plate, root pass 71 25 If diameter > 50 mm, stress concentration of cutout has
removed by drilling. to be considered
page 70

842 Nozzle welded on pipe, root pass as 63 22 If diameter > 50 mm, stress concentration of cutout has
welded. to be considered
900 Tubular joints
911 Circular hollow section butt joint to 63 22 Root of weld has to penetrate into the massive bar in
massive bar, as welded order to avoid a gap perpenticular to the stress
direction.
912 Circular hollow section welded to 63 22 Root of weld has to penetrate into the backing area in
component with single side butt weld, order to avoid a gap perpenticular to the stress
backing provided. direction.
Root crack.
page 71

913 Circular hollow section welded to 50 18 Impairment of inspection of root cracks by NDT may
component single sided butt weld or be compensated by adequate safety considerations (see
double fillet welds. chapter 5) or by downgrading down to 2 FAT classes.
Root crack.
921 Circular hollow section with welded on Non load-carrying weld.

disk
Fillet weld, toe ground 90 32
Fillet welds, as welded 71 25
931 Tube-plate joint, tubes flattened, butt 71 25-

weld (X-groove)
Tube diameter < 200 mm and

plate thickness < 20 mm
page 72

932 Tube-plate joint, tube slitted and

welded to plate
tube diameter < 200 mm and 63 18

plate thickness < 20 mm
tube diameter > 200 mm or 45 14
plate thickness > 20 mm
Tab. {3.2}-2: Fatigue resistance values for structural details on the basis of shear stress
No. Description FAT FAT

1 Parent metal or full penetration butt weld; m=5 down to 1E8 cycles 100 36
2 Fillet weld or partial penetration butt weld; m=5 down to 1E8 cycles 80 28
page 73
3.3 FATIGUE RESISTANCE AGAINST STRUCTURAL HOT

SPOT STRESS
3.3.1 Fatigue Resistance using Reference S-N Curve
The S-N curves for fatigue resistance against structural hot spot stress (2.2.3) are given in the
table {3.3}-1 for steel and aluminium, where the definition of the FAT class is given in chapter
3.2. The resistance values refer to the as-welded condition unless stated otherwise. The effects
of welding residual stress are included. Effects of misalignment are not included.
The design value of the structural hot spot stress range shall not exceed )Fhs < 2Afy.
Tab. {3.3}-1: Fatigue resistance against structural hot spot stress

No Structural detail Description Requirements FAT FAT
. Steel Alu.
1 Butt joint As welded, NDT 100 40
2 Cruciform or T-joint K-butt welds, no lamellar 100 40

with full penetration tearing
K-butt welds
3 Non load-carrying Transverse non-load 100 40

fillet welds carrying attachment, not
thicker than main plate, as
welded
4 Bracket ends, ends Fillet welds welded around 100 40

of longitudinal or not, as welded
stiffeners
5 Cover plate ends As welded 100 40

and similar joints
6 Cruciform joints Fillet welds, as welded 90 36

with load-carrying
fillet welds
page 74
No Structural detail Description Requirements FAT FAT

. Steel Alu.
7 Lap joint with load Fillet welds, as welded 90 36

carrying fillt welds
8 Type “b” joint with Fillet or full penetration 100 40

short attachment weld, as welded
9 Type “b” joint with Fillet or full penetration 90 36

long attachment weld, as welded
Note: Table does not cover effects of misalignment. They have to be considered explicitely in
determination of stress range, see also 3.8.2.
For hollow section joints, special hot-spot stress design S-N curves have been recommended
by the IIW [xx]. The tubular joint design curves should not be applied to other types of
structure.
3.3.2 Fatigue Resistance Using a Reference Detail
The tables of the fatigue resistance of structural details given in 3.2, or fatigue data from other
sources which refer to a comparable detail, may be used. The reference detail should be chosen
as similar as possible to the detail to be assessed. Thus the procedure will be:
a) Select a reference detail with known fatigue resistance, which is as similar as possible to
the detail being assessed with respect to geometric and loading parameters.
b) Identify the type of stress in which the fatigue resistance is expressed. This is usually
nominal stress (as in tables in chapter 3.2).
c) Establish a FEM model of the reference detail and the detail to be assessed with the same
type of meshing and elements following the recommendations given in 2.2.3.
d) Load the reference detail and the detail to be assessed with the stress identified in b).
e) Determine the structural hot spot stress Fhs, ref of the reference detail and the Structural
hot spot stress Fhs, assess of the detail to be assessed.
page 75
f) The fatigue resistance for 2 million cyles of the detail to be assessed FATassess is then
calculated from fatigue class of the reference detail FATref by:
page 76
3.4 FATIGUE RESISTANCE AGAINST EFFECTIVE

NOTCH STRESS
3.4.1 Steel
The effective notch stress fatigue resistance against fatigue actions, as determined in 2.2.4 for
steel, is given in table {3.4}-1. The definition of the FAT class is given in chapter 3.2. The fatigue
resistance value refers to the as-welded condition. The effect of welding residual stresses is
included. Possible misalignment is not included.
Tab. {3.4}-1: Effective notch fatigue resistance for steel
No. Quality of weld notch Description FAT
1 Effective notch radius Notch as-welded, normal 225

equalling 1 mm replacing welding quality
weld toe and weld root
notch m=3
3.4.2 Aluminium
Tab. {3.4}-2: Effective notch fatigue resistance for aluminium
No. Quality of weld notch Description FAT
1 Effective notch radius Notch as-welded, normal 75

equalling 1 mm replacing welding quality
weld toe and weld root
notch m=3
page 77
3.5 FATIGUE STRENGTH MODIFICATIONS

3.5.1 Stress Ratio
3.5.1.1 Steel
For stress ratios R<0.5 a fatigue enhancement factor f(R) may be considered by multiplying the
fatigue class of classified details by f(R). The fatigue enhancement factor depends on the level
and direction of residual stresses. It should only be used if reliable information or estimation of
the residual stress level was present. The following cases are to be distinguished:
I: Unwelded base material and wrought products with negligible residual stresses (<0.2Afy),
stress relieved welded components, in which the effects of constraints or secondary
stresses have been considered in analysis. No constraints in assembly.
f(R) = 1.6 for R < -1

f(R) = -0.4 @ R + 1.2 for -1 # R # 0.5
f(R) = 1 for R > 0.5
II: Small scale thin-walled simple structural elements containing short welds. Parts or
components containing thermally cut edges. No constraints in assembly.
f(R) = 1.3 for R < -1

f(R) = -0.4 @ R + 0.9 for -1 # R # -0.25
f(R) = 1 for R > -0.25
III: Complex two- or three-dimensional welded components, components with global residual
stresses, thickwalled components. Normal case for welded components and structures.
f(R) = 1 no enhancement
Fig. (3.5)-1 Enhancement factor f(R)

page 78
The ranking in categories I, II or III should be done and documented by the design office. If no
reliable information on residual stress is available, f(R)=1.
It has to be noted in this respect that stress relief in welded joints is unlikely to be fully effective,
and long range residual stresses may be introduced during assembly of prefabricated welded
components. For such reasons, it is recommended that values of f(R)>1 should only be adopted
for welded components in very special circumstances.
3.5.1.2 Aluminium
The same regulations as for steel are recommended.
3.5.2 Wall Thickness
3.5.2.1 Steel
The influence of plate thickness on fatigue strength should be taken into account in cases where
cracks start from the weld toe. The fatigue resistance values here given refer to a wll thickness
of 25 mm at steel. The reduced strength is taken in consideration by multiplying the fatigue class
of the structural detail by the thickness reduction factor f(t). The thickness correction exponent
n is dependent on the effective thickness teff and the joint category (see table {3.5}-1) [21]. The
same way a benign thinness correction might be applied down to 10 mm wall thickness.
Tab. {3.5}-1: Thickness correction exponents
Joint category Condition n
Cruciform joints, transverse T-joints, as-welded 0.3

plates with transverse attachments
Cruciform joints, transverse T-joints, toe ground 0.2
plates with transverse attachments
Transverse butt welds as-welded 0.2
Butt welds ground flush, base material, any 0.1
longitudinal welds or attachements
The plate thickness correction factor is not

required in the case of assessment based on
effective notch stress procedure or fracture
mechanics.
page 79
3.5.2.2 Aluminium
The same regulations as for steel are recommended.
Fig. (3.5)-2 Toe distance

3.5.3 Improvement Techniques
3.5.3.1 General
Post weld improvement techniques may raise the fatigue resistance. These techniques improve
the weld profile, the residual stress conditions or the environmental conditions of the welded
joint. The improvements methods are:
a) Methods of improvement of weld profile:
Machining or grinding of weld seam flush to surface

Machining or grinding of the weld transition at the toe
Remelting of the weld toe by TIG-, plasma or laser dressing
b) Methods for improvement of residual stress conditions:
Peening (hammer-, needle-, shot- or brush-peening)

Coining
Overstressing
Stress relieving thermal treatment
c) Methods for improvement of environmental conditions:
Painting
Resin coating
The effects of all improvement techniques are sensitive to the method of application and the
applied loading, being most effective in the low stress high cycle regime. They may also depend
on the material, structural detail and dimensions of the welded joint. Consequently, fatigue tests
for the verification of the procedure in the endurance range of interest are recommended
(chapters 3.7 and 4.5).
For some post welding improvement procedures, direct recommendations are given below. They
may be used onder the follwing circumstances:
a) Increasing the fatigue strength of new structures. A verification by test is

recommended
b) Repair or upgrading of existing structures
page 80
The recommendations apply to nominal stress and structural hot spot stress method, they do not
Fig: (3.5)-3: Examples of joint suitable for improvement

apply to effective notch stress and fracture mechanics method.
3.5.3.2 Applicabiliy of Improvement Methods
The recommendations apply to all arc welded steel or aluminium components subjected to
fluctuating or cyclic stress and designed to fatigue limit state criterion. They are limited to
structural steels up to a specified yield strength of 900 MPa and to structural aluminium alloys
commonly used in welded structures, primarily of the AA 5000 and AA 6000 series.
The recommendations apply to welded joints of plates, of sections built up of plates or similar
rolled or extruded shapes, and hollow sections. If not specified else, the plate thickness range for
steel is from 6 to 150 mm, for aluminium from 4 to 50 mm.
The application is limited to joints operating at temperatures below the creep range. In general,
the recommendations do not apply at low cycle fatigue conditions, so the nominal stress range
is limited to . For the special improvement procedures additional restrictions may
be given.
The improvement procedures described below, apply solely to the weld toe and to cracks starting
from this point. All other points of a possible start of fatigue cracks therefore should be carefully
considered as e.g. the weld root or weld imperfections.
page 81
The recomendations do not apply to joints operating under free corrosion.
Fig: (3.5)-4: Examples of joints, at which an improvement

might be limited by a possible root crack
3.5.3.3 Burr Grinding
Fatigue cracks initiate usually at the weld toe at points of cold fusion or other sharp crack-like
defects of a few tenth of a millimeter. The grinding has firstly to remove these defects and
secondly to create a smooth weld transition and thus to reduce the stress concentration. All
embedded imperfection which emerge to the surface at grinding must be repaired. For the details
of application see ref. [xx].
The benefit of burr grinding is given as a factor on the stress range of the fatigue class of a non-
improved joint.
Tab. 3.5-1: Benefit factors on stress of burr grinding

Area of application Mild steel Steel fy > 355 MPa
fy < 355 MPa & aluminium alloys
All structural details leading to a IIW fatigue class 1.3 1.5
of 90 at steel or 32 at aluminium or lower as (1.3*)
applicable
All S-N curves and fatigue classes for assessment 1.3 1.5
by structural hot spot stress, but not higher than (1.3*)
100 at steel and 40 at aluminium. Butt joints to be
assessed by the nominal stress fatigue class.
*
For transverse fillet welds at continous plates, corresponding to FAT 80 at steel and FAT 28 at
aluminium in the catalogue of details
page 82
Fig. (3.5)-5: Benefit of

burr grinding at steel
Fig. (3.5)-6: Benefit of

burr grinding at aluminium
page 83
3.5.3.4 TIG Dressing
By TIG (tungsten inert gas) dressing, the weld toe is remolten in order to remove the weld toe
undercut or other irregularities and to smoothen the stress concentration of the weld transition.
The recommendations apply to partial or full penetration arc welded fillet welds in steels with a
specified yield strength up to 900 MPa and to wall thicknesses >= 10 mm operating non-
corrosive environment or under conditions of corrosion potection. The details of the procedure
are described in ref. [xx].
Tab. 3.5-2: Benefit factors on stress of TIG dressing

Area of application Mild steel Steel fy > 355 MPa
fy < 355 MPa & aluminium alloys
All structural details leading to a IIW fatigue 1.3 1.5
class of 90 at steel or lower (1.3*)
All S-N curves and fatigue classes for 1.3 1.5
assessment by structural hot spot stress, but (1.3*)
not higher than 100 at steel and 40 at
aluminium. Butt joints to be assessed by the
nominal stress fatigue class.
*
For transverse fillet welds at continous plates, corresponding to FAT 80 at steel and FAT 28 at
aluminium in the catalogue of details
Fig. (3.5)-7: Benefits of by

TIG dressing at steel
Fig. (3.5)-8: Benefit of TIG

dressing at aluminium
page 84
3.5.3.5 Hammer Peening
By hammer peening, the material is plastically deformed at the weld toe in order to introduce
beneficial compressive residual stresses. The recommendation is restricted to steels with a
specified yield strength up to 900 MPa and structural aluminium alloys, both operating non-
corrosive environment or under conditions of corrosion potection. The recommendations apply
for plate thicknesses from 10 to 50 mm at steel and 5 to 25 mm at aluminium (?) and to arc
welded fillet welds with a minumum weld leg length of 0.1@t, where t is the wall thickness of the
stressed plate. The details of the procedure are described in ref. [xx].
Tab. 3.5-3: Benefit on stress of hammer peening (nominal stress)

Area of application Benefit Requirements
All structural details Upgrade for steel Max. amount of nominal compressive stress
leading to a IIW to FAT 125 in load spectrum , including
fatigue class of 90 at for aluminium to proof loading
steel or 40 at FAT 56 if R<0 then use
aluminium or lower if R>=0 then use instead of
(for aluminium use fy of HAZ!)
For structural hot spot stress see recommendations for needle peening
Fig. (3.5)-9: Benefits of

hammer peening at steel
page 85
Fig. (3.5)-10: Benefits of improve-

ment by hammer peening at alumi-
nium
3.5.3.6 Needle Peening
By needle peening, the material is plastically deformed at the weld toe in order to introduce
beneficial compressive residual stresses. Before any application, it is recommended to grind the
weld toe in order to remove undercut and weld toe irregularities and subsequently to finish with
a sandpaper tool for a glossy surface. The details of the procedure are described in [xx].
Tab. 3.5-4: Benefit on stress of needle peening (nominal stress)

Area of application Benefit Requirements
All structural Upgrade for steel to Max. amount of nominal compressive
details leading to a FAT 125 stress in load spectrum
IIW fatigue class for aluminium to if R<0 then use
of 90 at steel or 40 FAT 56 if R>=0 then use instead of
at aluminium or
(for aluminium use fy of HAZ!)
lower
Fig. (3.5)-12: Benenfits of

needle peening at aluminium
page 86
At all peening techniques, the structural hot spot stress approach should be applied only to joints
with fillet welds (with any penetration) and not to butt joints. The structural hot spot stress,
which includes the stress increase due to the structural geometry and possible misalignments can
be assessed by the corresponding material S-N curve, e.g. FAT 160 for steel and FAT 71 for
aluminium alloys in conjunction with the slope exponent m=5.0 . In this way, the base metal at
the weld toe is assumed to have a lower fatigue strength than the peened weld.
Fig. (3.5)-11: Benefits of by

needle peening at steel
3.5.4 Effect of Elevated Temperatures
3.5.4.1 Steel
For higher temperatures, the fatigue resistance data may be modified with a reduction factor
given in fig. (3.5)-13. The fatigue reduction factor is a conservative approach and might be raised
according to test evidence or application codes.
page 87
Fig. (3.5)-13 Fatigue strength reduction factor for steel at elevated temperatures
3.5.4.2 Aluminium
The fatigue data given here refer to operation temperatures lower than 70 °C. This value is a
conservative approach. It may be raised according to test evidence or an applicable code.
3.5.5 Effect of Corrosion
The fatigue resistance data given here refer to non-corrosive environments. Normal protection
against atmospheric corrosion is assumed. A corrosive environment or unprotected exposure to
atmospheric conditions may reduce the fatigue class. The fatigue limit may also be reduced con-
siderably. The effect depends on the spectrum of fatigue actions and on the time of exposure.
For steel, except stainless steel, in marine environment not more than 70% of the fatigue
resistance values in terms of stress range shall be applied and no fatigue limit be considered. In
fracture mechanics crack propagation calculations the constant C0 of the Paris Power Law shall
be multiplied by a factor of 3.0 . A threshold value shall not be considered.
No further specific recommendations are given for corrosion fatigue assessment.
page 88
3.6 FATIGUE RESISTANCE AGAINST CRACK PRO-

PAGATION
The resistance of a material against cyclic crack propagation is characterized by the material
parameters of the "Paris" power law of crack propagation
where the material parameters are
C0 constant of the power law

m exponent of the power law
)K range of cyclic stress intensity factor
)Kth threshold value of stress intensity
R ratio Kmin/Kmax, taking all stresses including residual stresses into account (see
3.5.1)
In the absence of specified or measured material parameters, the values given below are
recommended. They are characteristic values.
3.6.1 Steel
(units in MPa%m and m) or
check numbers!!
C0 = 1.58 @10-11
C0 = 5.0 A10-13 (units in N*mm-3/2 and mm)
m =3
)Kth = 6.0 - 4.56AR but not lower than 2 (units in MPa%m) or

)Kth = 190 - 144 AR but not lower than 62 (units in N*mm-3/2)
3.6.2 Aluminium
C0 = 1.27 @10-9 (units in MPa%m and m) or
C0 = 4.00 A10-11 (units in N*mm-3/2 and mm)
m =3
)Kth = 2.0 - 1.5 AR but not lower than 0.7 (units in MPa%m) or
)Kth = 63 - 48 AR but not lower than 21 (units in N*mm-3/2)
page 89
3.7 FATIGUE RESISTANCE DETERMINATION BY

TESTING
3.7.1 General Considerations
Fatigue tests may be used to establish a fatigue resistance curve for a component or a structural
detail, or the resistance of a material against (non critical) cyclic crack propagation.
Statistical methods offer three ways of testing a limited number of samples from a larger
population:
1. a specimen to failure
2. first specimen to failute
3. p specimens to failure amongst n specimens
It is recommended that test results are obtained at constant stress ratios R. The S-N data should
be presented in a graph showing log(endurance in cycles) as the abscissa and log(range of fatigue
actions) as the ordinate. For crack propagation data, the log(stress intensity factor range) should
be the abscissa and the log(crack propagation rate per cycle) the ordinate.
Experimental fatigue data are scattered, the extent of scatter tends to be greatest in the low
stress/low crack propagation regime (e.g. see fig. (3.7)-1). For statistical evaluation, a Gaussian
log-normal distribution should be assumed. The number of failed test specimens should be equal
or greater than 10. For other conditions, special
statistical considerations are required.
Many methods of statistical evaluation are

available. However, the most common approach
for analysing fatigue data is to fit S-N or crack
propagation curves by regression analysis, taking
log(N) or log(da/dN) as the dependent variable.
Then, characteristic values are established by
adopting curves lying k standard deviations of
the dependent variable from the mean (values are
given in 6.4.1). In the case of S-N data, this Fig. (3.7)-1 Scatterband in SN curve
would be below the mean, while the curve above
the mean would be appropriate in case of crack
propagation data.
Thus, more precisely, test results should analysed to produce characteristic values (subscript k).
These are values at a 95% survival probability (i.e. 5% failure probability) associated with a two-
sided 75% confidence level of the mean.
More details on the use of the confidence level and formulae are given in appendic 6.4.
page 90
3.7.2 Evaluation of Test Data

Different methods for testing exist. For the derivation of S-N curves, testing at two levels of the
stress range )F within the range of 105 to 106 cycles is preferred. For fracture mechanics crack
propagation parameters, the range of stress intensity factor )K should be distributed between
threshold and brittle fracture levels.
For the evaluation of test data originating from a test series, the characteristic values are
calculated by the following procedure:
a) Calculate log10 of all data: Stress range )F and number of cycles N, or stress
intensity factor range )K and crack propagation rate da/dN.
b) Calculate exponents m and constant logC (or logC0 resp.) of the formulae:
by linear regression taking stress or stress intensity factor range as the

independent variable, i.e. logN=f(log )F) or log(da/dN)=f(log )K).
If the number of pairs of test data n<10, or if the data are not sufficiently evenly
distributed to determine m accurately, a fixed value of m should be taken, as
derived from other tests under comparable conditions, e.g. m=3 for steel and
alumunium welded joints.
If a fixed value of m is used, the values xi equalling logC or logC0 are calculated
from the (N, )F)i or (N, da/dN)i test results using the equations above.
c) Calculate mean xm and standard deviation Stdv of logC (or logC0 resp.) using m
obtained in b).
d) Calculate the characteristc values xk by the formula
The values of k are given in table {3.7}-1.
page 91
Tab. {3.7}-1: Values of k for the calculation of characteristic values
n 10 15 20 25 30 40 50 100
k 2.7 2.4 2.3 2.2 2.15 2.05 2.0 1.9
For n<10 and more details and information, see appendix 6.4.1 and ref. [35, xx].
In case of S-N data, proper account should be taken of the fact that residual stresses are usually
low in small-scale specimens. The results should be corrected to allow for the greater effects of
residual stresses in real components and structures. This may be achieved either by testing at high
R-ratios, e.g. R=0.5, or by testing at R=0 and lowering the fatigue strength at 2 million cycles
(FAT) by 20% .
3.7.3 Evaluation of Data Collections
Usually data collections do not origin from a single statistical population. These heterogeneous
populations of data require a special consideration in order to avoid an excessive and unnecessary
calculative scatter. The evaluation procedure should consist of the following steps:
1. Calculate the constant C of the SN Wöhler curve for each data point (eq.1) using the
anticipated knowledge of the slope exponent of comparable test series, e.g slope m=3.00
for steel or aluminium.
2. Plot all values C into a Gaussian probability chart, showing the values of C on the
abscissa and the cumulative survival probability on the ordinate.
3. Check the probability plot for heterogeneity of the population. If it is heterogeneous,
separate the portion of the population which is of interest (see illustration on figures
(3.7)-2 and (3.7-3)).
4. Evaluate the interesting portion of population according to chapter 3.7.2.
Fig. (3.7)-2: Example of scatter in data collections
page 92
Fig. (3.7)-3: Example of a heterogeneous

population
page 93
3.8 FATIGUE RESISTANCE OF JOINTS WITH WELD

IMPERFECTIONS
3.8.1 General
3.8.1.1 Types of Imperfections
The types of imperfections covered in this document are listed below. Other imperfections, not
yet covered, may be assessed by assuming similar imperfections with comparable notch effect.
Imperfect shape
All types of misalignment including centre-line mismatch (linear misalignment) and

angular misalignment (angular distortions, roofing, peaking).
Undercut
Volumetric discontinuities
Gas pores and cavities of any shape.
Solid inclusions, such as isolated slag, slag lines, flux, oxides and metallic inclusions.
Planar discontinuities
All types of cracks or cracklike imperfections, such as lack of fusion or lack of

penetration (Note that for certain structural details intentional lack of penetration is
already covered, e.g. at partial penetration butt welds or cruciform joints with fillet
welds)
If a volumetric discontinuity is surface breaking or near the surface, or if there is any

doubt about the type of an embedded discontinuity, it shall be assessed like a planar
discontinuity.
3.8.1.2 Effects and Assessment of Imperfections
At geometrical imperfections, three effects affecting fatigue resistance can be distiguished, as

summarized in table {3.8}-1.
page 94
Increase of general stress level
This is the effect of all types of misalignment due to secondary bending. The additional
effective stress concentration factor can be calculated by appropriate formulae. The
fatigue resistance of the structural detail under consideration is to be lowered by division
by this factor.
Local notch effect
Here, interaction with other notches present in the welded joint is decisive. Two cases are
to be distinguished:
Additive notch effect
If the location of the notch due to the the weld imperfection coincides with a structural
discontinuity associated with the geometry of the weld shape (e.g. weld toe), then the
fatigue resistance of the welded joint is decreased by the additive notch effect. This may
be the case at weld shape imperfections.
Competitive notch effect
If the location of the notch due to the weld imperfection does not coincide with a
structural geometry associated with the shape geometry of the weld, the notches are in
competition. Both notches are assessed separately. The notch giving the lowest fatigue
resistance is governing.
Cracklike imperfections
Planar discontinuities, such as cracks or cracklike imperfections, which require only a

short period for crack initiation, are assessed using fracture mechanics on the basis that
their fatigue lives consist entirely of crack propagation.
After inspection and detection of a weld imperfection, the first step of the assessment procedure
is to determine the type and the effect of the imperfection as given here.
If a weld imperfection cannot be clearly associated to a type or an effect of imperfections listed

here, it is recommended that it is assumed to be cracklike.
page 95
Tab. {3.8}-1: Categorization and assessment procedure for weld imperfections
Effect of imperfection Type of imperfection Assessment
Rise of general stress Misalignment Formulae for effective

level stress concentration
Local additive Weld shape imperfections, Tables given
notch undercut
effect
competitive Porosity and inclusions Tables given
not near the surface
Cracklike imperfection Cracks, lack of fusion and Fracture mechanics
penetration, all types of
imperfections other than
given here
3.8.2 Misalignment
Misalignment in axially loaded joints leads to an increase of stress in the welded joint due to the
occurrence of secondary shell bending stresses. The resulting stress is calculated by stress
analysis or by using the formulae for the stress magnification factor km given in appendix 6.3.
Secondary shell bending stresses do not occur in continuous welds longitudinally loaded or in
joints loaded in pure bending, and so misalignment will not reduce the fatigue resistance.
However, misalignment in components, e.g. beams, subject to overall bending may cause
secondary bending stresses in parts of the component, where the through thickness stress
gradient is small, e.g. in a flange of a beam, where the stress is effectively axial. Such cases
should be assessed.
Some allowance for misalignment is already included in the tables of classified structural details
(3.2). In particular, the data for transverse butt welds are directly applicable for misalignment
which results in an increase of stress up to 30%, while for the cruciform joints the increase can
be up to 45% . In local concepts of fatigue analysis, a small but unevitable amount of
misalignment according to a stress manification factor of km =1.05 is already included in the
fatigue resistance S-N curves.
In special joints, i.e. all listed in table 3.8-2, the effect of a larger misalignment has to be
additionally considered in the local stress (structural hot spot stress or effective notch stress. The
misalignement effect may be present even in the vicinity of supporting structures. A
corresponding stress increase has to be taken into account also in crack propagation analyses. In
all those cases, where the stress magnification factor is directly calculated, the effective stress
magnification factor km, eff should be calculated as given below.
page 96
For the simultaneous occurrence of linear and angular misalignment, both stress magnification
factors should be applied simultaneously using the formula:
As misalignment reduces the fatigue resistance, the fatigue resistance of the classified structural
detail (3.2) has to be divided by the effective stress magnification factor.
Tab. {3.8}-2: Consideration of stress magnification factors due to misalignment
Type of km Nominal stress Structural hot spot and effective notch approach
analysis approach
Type of welded km already co- km already covered Default value of

joint vered in FAT in SN curves effective km to be
class considered in stress
Butt joint made 1.15 1.05 1.10*

in shop in flat
position
Other butt 1.30 1.05 1.25*
joints
cruciform joints 1.45 1.05 1.40*
One-sided fillet 1.25 1.05 1.20**

welds
*) but not more than (0.95 + 3A emax /t), where emax = permissible misalignment and t =
wall thickness of loaded plate
**) but not more than (0.95 + 0.3A tref/t), where tref = reference wall thickness
page 97
3.8.3 Undercut
The basis for the assessment of undercut is the ratio u/t, i.e. depth of undercut to plate thickness.
Though undercut is an additive notch, it is already considered to a limited extent in the tables of
fatigue resistance of classified structural details (3.2).
Undercut does not reduce fatigue resistance of welds which are only longitudinally loaded.
3.8.3.1 Steel
Tab. {3.8}-3: Acceptance levels for weld toe undercut in steel
Fatigue class Allowable undercut u/t

butt welds fillet welds
100 0.025 not applicable

80 0.075 0.05
71 0.10 0.075
63 0.10 0.10
56 and lower 0.10 0.10
Notes: a) undercut deeper than 1 mm shall be assessed as a
crack-like imperfection.
b) the table is only applicable for plate thick-
nesses from 10 to 20 mm
3.8.3.2 Aluminium
Tab. {3.8}-4: Acceptance levels for weld toe undercut in aluminium
Fatigue class Allowable undercut u/t

butt welds fillet welds

40 0.075 0.05
36 0.10 0.075
32 0.10 0.10
28 and lower 0.10 0.10
Notes: a) undercut deeper than 1 mm shall be assessed as a
crack-like imperfection.
b) the table is only applicable for plate thick-
nesses from 10 to 20 mm
page 98
3.8.4 Porosity and Inclusions

Embedded volumetric discontinuities, such as porosity and inclusions, are considered as
competitive weld imperfections which can provide alternative sites for fatigue crack initiation
than those covered by the fatigue resistance tables of classified structural details (3.2).
Before assessing the imperfections with respect to fatigue, it should be verified that the
conditions apply for competitive notches, i.e. that the anticipated sites of crack initiation in the
fatigue resistance tables do not coincide with the porosity and inclusions to be assessed and no
interaction is expected.
It is important to ensure that there is no interaction between multiple weld imperfections, be it

from the same or different type. Combined porosity or inclusions shall be treated as a single large
one. The defect interaction criteria given in (3.8.5) for the assessment of cracks also apply for
adjacent inclusions. Worm holes shall be assessed as slag inclusions.
If there is any doubt about the coalescence of porosity or inclusions in the wall thickness
direction or about the distance from the surface, the imperfections shall be assessed as cracks. It
has to be verified by NDT that the porosity or inclusions are embedded and volumetric. If there
is any doubt, they are to be treated as cracks.
The parameter for assessing porosity is the maximum percentage of projected area of porosity in
the radiograph; for inclusions, it is the maximum length. Directly adjacent inclusions are regarded
as a single one.
3.8.4.1 Steel
Tab. {3.8}-5: Acceptance levels for porosity and inclusions in welds in steel
Fatigue class Max. length of an inclusion in mm Limits of

porosity in %
as-welded stress relieved + of area * **
100 1.5 7.5 3
90 2.5 19 3
80 4 58 3
71 10 no limit 5
63 35 no limit 5
56 and lower no limit no limit 5
* Area of radiograph used is length of weld affected by porosity
multiplied by width of weld
** Maximum pore diameter or width of an inclusion less than 1/4
plate thickness or 6 mm
+ Stress relieved by post weld heat treatment
page 99
3.8.4.2 Aluminium
Tab. {3.8}-6: Acceptance levels for porosity and inclusions in welds in aluminium
Fatigue class Max. length of an Limits of porosity

inclusion in mm ** in % of area * **
as-welded
40 and higher 1.5 0 +)

36 2.5 3
32 4 3
28 10 5
25 35 5
15 and lower no limit 5
* Area of radiograph used is length of weld affected by
porosity multiplied by width of weld
** Maximum pore diameter or width of an inclusion less
than 1/4 plate thickness or 6 mm
+) Single pores up to 1.5 mm allowed
Tungsten inclusions have no effect on fatigue behaviour and therefore do not need to be
assessed.
page 100
3.8.5 Cracklike Imperfections
3.8.5.1 General Procedure
Planar discontinuities, cracks or cracklike defects are identified by non-destructive testing and
inspection. NDT indications are idealized as elliptical cracks for which the stress intensity factor
is calculated according to 2.2.5.
Fig. (3.8)-1 Transformation of NDT indications to an elliptic or semi-elliptic
For embedded cracks, the shape is idealized by a circumscribing ellipse, which is measured by
its two half-axes a and c. The crack parameter a (crack depth) is the half-axis of the ellipse in the
direction of the crack growth to be assessed. The remaining perpendicular half-axis is the half
length of the crack c. The wall thickness parameter t is the distance from the center of the ellipse
to the nearest surface. If the ratio a/t > 0.75, the defect is to be recategorized as a surface defect.
Surface cracks are described in terms

of a circumscribing halfellipse. The wall
thickness parameter is wall thickness t.
If the ratio of a/t>0.75, the defect is
regarded as being fully penetrating and
is to be recategorized as a centre crack
or an edge crack, whichever is
applicable.
Fig. (3.8)-2 Crack dimensions for assessment
For details of dimensions of cracks and
recategorization see appendix 6.2.
3.8.5.2 Simplified Procedure
The simplified procedure is based on the integration of the crack propagation law (4.4) from an
initial defect size ai to defect size of 0.75% of wall thickness using the material resistance against
crack propagation as given in 3.6.1 for steel. For cracks near the plate edge, the distance b from
the center of crack ellipsis to the plate edge was constantly assumed equalling c. This ensures
page 101
conservative results.
In the tables the stress ranges at 2*106 cycles corresponding to the definition of the fatigue
classes (FAT) of classified structural details (3.2) are shown. The tables have been calculated
using the correction functions and the weld joint local geometry correction given in 6.2.4. (see
tab. {6.2}-1 and tab. {6.2}-3).
In assessing a defect by the simplified procedure, the stress range )Fi for the initial crack size
parameter ai and the stress range )Fc for the critical crack size parameter ac are taken. The stress
range )F or the FAT class belonging to a crack propagation from ai to ac at 2A106 cycles is then
calculated by:
For aluminium, the tables may be used by dividing the resistance stress ranges at 2A 106 cycles
(FAT classes) for steel by 3.
Fig. (3.8)-3 Toe distance l for simplified

procedure
Tables {3.8}-7: Stress ranges at 2A106 cycles (FAT classes in N/mm2) of welds containing cracks
for the simplified procedure (following 3 pages)
page 102
Surface cracks at fillet weld toes
ai long surface crack near plate edge, fillet welds l/t=2.5 a/c=0.1
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 7 16
20.0 0 0 0 0 0 0 0 0 0 0 0 4 6 8 11 19
16.0 0 0 0 0 0 0 0 0 0 0 4 7 9 11 15 22
12.0 0 0 0 0 0 0 0 0 0 6 9 12 14 16 19 25
10.0 0 0 0 0 0 0 0 0 5 9 12 15 17 19 22 27
8.0 0 0 0 0 0 0 4 7 9 13 16 19 21 22 25 30
6.0 0 0 0 0 0 6 9 12 15 18 21 23 25 26 28 33
5.0 0 0 0 0 6 10 13 16 18 21 24 26 28 29 31 35
4.0 0 0 0 5 10 14 18 20 22 25 28 29 31 32 33 37
3.0 0 0 7 11 16 20 23 25 27 30 32 33 34 35 37 39
2.0 5 11 16 20 25 28 31 32 34 36 37 39 40 40 41 43
1.0 22 28 32 34 38 40 42 43 44 45 46 47 48 48 48 48
0.5 38 42 45 47 49 51 52 53 53 54 54 54 54 54 54 52
0.2 57 59 61 61 63 63 63 63 63 63 63 62 61 61 60 56
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai long surface crack apart from edge, fillet welds l/t=2.5 a/c=0.1
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 13 22
20.0 0 0 0 0 0 0 0 0 0 0 0 7 11 13 17 25
16.0 0 0 0 0 0 0 0 0 0 0 9 12 15 18 21 28
12.0 0 0 0 0 0 0 0 0 0 11 15 18 21 23 26 32
10.0 0 0 0 0 0 0 0 0 10 15 19 22 24 26 28 34
8.0 0 0 0 0 0 0 8 12 15 20 24 26 28 29 32 37
6.0 0 0 0 0 0 12 16 19 22 26 29 31 33 34 36 40
5.0 0 0 0 0 11 17 20 24 26 29 32 34 35 36 38 42
4.0 0 0 0 9 17 22 26 28 30 33 36 37 39 40 41 44
3.0 0 0 13 18 25 29 32 34 36 38 40 42 43 44 45 47
2.0 11 19 25 29 34 37 39 41 42 44 46 47 48 49 50 51
1.0 32 38 42 44 48 50 52 53 54 55 56 57 57 57 57 56
0.5 50 53 56 58 60 62 63 63 64 64 64 64 63 63 62 59
0.2 70 72 73 74 75 75 74 74 74 73 72 71 70 69 67 62
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai short surface crack apart from edge, fillet welds l/t=2.5 a/c=.5
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 23 35
20.0 0 0 0 0 0 0 0 0 0 0 0 15 21 24 29 38
16.0 0 0 0 0 0 0 0 0 0 0 18 23 27 30 34 42
12.0 0 0 0 0 0 0 0 0 0 21 27 32 35 37 40 45
10.0 0 0 0 0 0 0 0 0 20 27 33 36 39 41 43 47
8.0 0 0 0 0 0 0 18 24 28 34 39 41 43 45 47 49
6.0 0 0 0 0 0 23 30 34 37 42 45 47 48 49 51 52
5.0 0 0 0 0 22 31 36 39 42 46 48 50 51 52 53 53
4.0 0 0 0 20 32 38 42 45 47 50 52 54 54 55 55 55
3.0 0 0 26 33 42 47 50 52 53 55 57 58 58 58 58 57
2.0 22 36 43 48 53 56 58 60 61 62 62 62 62 62 62 59
1.0 53 60 63 66 68 69 70 70 70 70 69 69 68 67 66 62
0.5 74 76 78 78 79 78 78 77 77 76 74 73 72 71 69 64
0.2 92 91 91 90 88 86 85 84 83 81 79 77 75 74 72 65
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
page 103
Surface cracks at butt weld toes
ai long surface crack near plate edge, butt welds l/t=1 a/c=0.1
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 7 17
20.0 0 0 0 0 0 0 0 0 0 0 0 4 6 8 11 20
16.0 0 0 0 0 0 0 0 0 0 0 4 7 9 11 15 23
12.0 0 0 0 0 0 0 0 0 0 6 9 12 14 16 20 27
10.0 0 0 0 0 0 0 0 0 5 9 12 15 18 20 23 29
8.0 0 0 0 0 0 0 4 7 9 13 17 19 22 23 26 32
6.0 0 0 0 0 0 6 9 12 15 18 22 25 26 28 30 35
5.0 0 0 0 0 6 10 13 16 18 22 25 28 29 31 33 38
4.0 0 0 0 5 10 14 18 21 23 26 29 31 33 34 36 40
3.0 0 0 7 11 16 21 24 27 29 31 34 36 37 38 40 43
2.0 5 11 16 20 26 30 32 34 36 38 40 42 43 44 45 48
1.0 22 29 33 36 40 43 45 46 47 49 51 52 52 53 53 54
0.5 41 45 48 50 53 55 57 58 58 59 60 60 60 60 60 59
0.2 61 64 66 68 69 70 70 70 70 70 70 70 69 69 68 64
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai long surface crack apart from edge, butt welds l/t=1 a/c=0.1
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 13 23
20.0 0 0 0 0 0 0 0 0 0 0 0 7 11 13 17 26
16.0 0 0 0 0 0 0 0 0 0 0 9 12 15 18 21 29
12.0 0 0 0 0 0 0 0 0 0 11 15 18 21 23 26 33
10.0 0 0 0 0 0 0 0 0 10 15 19 22 25 27 29 36
8.0 0 0 0 0 0 0 8 12 15 20 24 27 29 31 33 39
6.0 0 0 0 0 0 12 16 19 22 26 30 32 34 35 37 43
5.0 0 0 0 0 11 17 20 24 26 30 33 35 37 38 40 45
4.0 0 0 0 9 17 22 26 29 31 34 37 39 41 42 44 48
3.0 0 0 13 18 25 29 32 35 37 40 42 44 45 46 48 51
2.0 11 19 25 29 35 38 41 43 45 47 49 50 51 52 53 55
1.0 33 39 43 46 50 53 54 56 57 59 60 61 61 61 62 61
0.5 52 56 59 61 64 66 67 68 68 69 69 69 69 69 68 66
0.2 74 77 78 79 80 81 81 80 80 80 79 78 77 76 75 70
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai short surface crack apart from edge, butt welds l/t=1 a/c=0.5
25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 23 36
20.0 0 0 0 0 0 0 0 0 0 0 0 15 21 24 29 40
16.0 0 0 0 0 0 0 0 0 0 0 18 23 27 30 34 43
12.0 0 0 0 0 0 0 0 0 0 21 27 32 35 37 41 47
10.0 0 0 0 0 0 0 0 0 20 27 33 37 39 41 44 50
8.0 0 0 0 0 0 0 18 24 28 34 39 42 44 46 48 52
6.0 0 0 0 0 0 23 30 34 37 42 46 48 50 51 53 56
5.0 0 0 0 0 22 31 36 39 42 47 50 52 53 54 56 57
4.0 0 0 0 20 32 38 43 46 48 52 54 56 57 58 59 60
3.0 0 0 26 33 42 47 51 53 55 58 59 61 61 62 62 62
2.0 22 36 43 48 54 58 60 62 63 65 66 67 67 67 67 65
1.0 54 61 65 68 71 73 74 74 75 75 75 75 74 74 73 69
0.5 76 80 82 83 84 84 84 84 84 83 82 81 80 79 77 71
0.2 98 98 98 98 97 95 94 93 92 90 88 86 85 83 81 74
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
page 104
Embedded cracks
ai embedded long crack near plate edge a/c=0.1

25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 7 17
20.0 0 0 0 0 0 0 0 0 0 0 0 3 5 7 11 20
16.0 0 0 0 0 0 0 0 0 0 0 4 6 9 11 14 24
12.0 0 0 0 0 0 0 0 0 0 5 8 11 14 16 19 28
10.0 0 0 0 0 0 0 0 0 4 8 12 15 17 19 23 31
8.0 0 0 0 0 0 0 3 6 8 12 16 19 22 24 27 35
6.0 0 0 0 0 0 5 9 12 14 18 22 25 27 29 32 40
5.0 0 0 0 0 5 9 12 15 18 22 26 28 31 32 35 43
4.0 0 0 0 4 9 14 17 20 23 26 30 33 35 37 39 47
3.0 0 0 6 10 16 20 24 27 29 32 36 38 40 42 45 52
2.0 4 10 15 19 25 30 33 36 38 41 44 46 48 50 52 59
1.0 22 29 33 37 42 46 49 51 53 56 59 61 63 64 67 73
0.5 42 47 52 55 60 63 66 68 69 72 75 77 79 80 82 88
0.2 68 73 77 80 84 87 90 92 93 96 98 100 101 103 105 110
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai embedded long crack apart from plate edge a/c=0.1

25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 19 30
20.0 0 0 0 0 0 0 0 0 0 0 0 12 17 20 24 33
16.0 0 0 0 0 0 0 0 0 0 0 14 19 22 25 29 37
12.0 0 0 0 0 0 0 0 0 0 17 22 26 29 31 34 41
10.0 0 0 0 0 0 0 0 0 16 22 27 30 33 35 37 44
8.0 0 0 0 0 0 0 14 19 23 28 32 35 37 39 41 48
6.0 0 0 0 0 0 19 24 28 31 35 38 41 42 44 46 53
5.0 0 0 0 0 18 25 29 33 35 39 42 44 46 47 49 56
4.0 0 0 0 15 26 31 35 38 40 43 46 48 50 51 54 59
3.0 0 0 21 27 35 39 42 45 46 49 52 54 56 57 59 64
2.0 17 29 35 40 45 49 51 53 55 58 60 62 64 65 67 71
1.0 44 51 55 58 62 65 67 69 70 73 75 77 78 79 80 84
0.5 65 69 73 75 79 82 84 85 86 88 90 91 92 93 94 97
0.2 91 95 98 100 103 105 106 107 108 110 111 112 112 113 114 117
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
ai embedded short crack apart from plate edge a/c=0.5

25.0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 27 41
20.0 0 0 0 0 0 0 0 0 0 0 0 17 24 28 34 46
16.0 0 0 0 0 0 0 0 0 0 0 20 27 32 35 40 50
12.0 0 0 0 0 0 0 0 0 0 24 32 37 41 43 47 55
10.0 0 0 0 0 0 0 0 0 23 32 38 42 45 48 51 58
8.0 0 0 0 0 0 0 20 28 33 40 45 48 51 53 56 62
6.0 0 0 0 0 0 27 35 40 43 48 53 56 58 59 62 67
5.0 0 0 0 0 26 36 42 46 49 54 57 60 62 63 66 71
4.0 0 0 0 23 37 44 49 53 56 60 63 65 67 68 70 75
3.0 0 0 31 39 49 54 58 61 64 67 70 71 73 74 76 80
2.0 26 42 50 55 62 67 70 72 74 76 79 80 81 82 84 87
1.0 62 70 75 78 83 86 88 89 91 92 94 95 96 97 98 100
0.5 88 93 96 99 102 104 105 106 107 108 110 110 111 112 112 115
0.2 118 121 123 124 126 128 129 129 130 131 132 133 133 134 135 137
t = 3 4 5 6 8 10 12 14 16 20 25 30 35 40 50 100
page 105
4 FATIGUE ASSESSMENT
4.1 GENERAL PRINCIPLES
In fatigue assessment, the fatigue actions and the fatigue resistance are related by means of
an appropriate assessment procedure. It must be ensured that all three elements (actions,
resistance and assessment procedure) correspond. Three procedures may be distinguished:
a) Procedures based on S-N curves, such as
nominal stress approach

Structural hot spot stress approach
effective notch stress approach
b) Procedures based on crack propagation considerations
c) Direct experimental approach by fatigue testing of components or entire

structures
4.2 COMBINATION OF NORMAL AND SHEAR

STRESS
If normal and shear stress occur simultaneously, their combined effect shall be considered.
Three cases may be distinguished:
a) If the equivalent nominal shear stress range is less than 15% of the
equivalent normal stress range or if the damage sum due to shear stress
range is lower than 10% of that due to normal stress range, the effect of
shear stress may be neclected.
b) If the normal and shear stress vary simultaneously in phase, or if the plane
of maximum principal stress is not changed significantly, the maximum
principal stress range may be used.
c) If normal and shear stress vary independently out of phase, in damage

calculation the damage sums shall be calculated separately and finally
added. The usage of 1/2 of the calculated life cycles is recommended.
Fracture mechanics crack propagation calculations should be based on maximum principal

stress range.
page 106
4.3 FATIGUE ASSESSMENT USING S-N CURVES

Fatigue verification is carried out using
the design spectrum of fatigue actions in terms of stress ranges )Fi,S,d, in which the
stresses of the characteristic spectrum )Fi,S,k have been multiplied by the partial
safety factor (F for fatigue actions.
and
the design resistance S-N curve based on design resistance stresses )FR,d, in which
the characteristic resistance stress ranges )FR,k have been divided by the partial
safety factor (M for fatigue resistance.
The design resistance S-N curve may be modified further according to the needs of the
damage calculation procedure.
For constant amplitude loading, the characteristic stress range )FR,k at the required
number of stress cycles is firstly determined. Secondly the fatigue criterion is checked:
At variable amplitude loading, cumulative damage calculation procedure is applied.

Usually a modified "Palmgren-Miner"-rule, as described in 4.3.1, is appropriate. For load
spectra which are sensitive to the position of the fatigue limit or cut-off limit, or in which
the spectrum changes during the service time, additional assessment using the nonlinear
damage calculation method described in 4.3.2 is recommended.
In fields of application, where no test data nor service experience exist and the shape of the
stress spectrum is not close to constant amplitude, it is recommended to proceed according
to the calculation given in in 4.3.1.
4.3.1 Linear Damage Calculation by "Palmgren-Miner" Summation
If the maximum design stress range )Fmax,S,d of the load spectrum is lower than the design
fatigue limit )FL,R,d of the design fatigue resistance S-N curve, or if it is lower than the
design cut-off limit )Fcut, R,d in cases where no fatigue limit is given, the life of the welded
joint can be assumed to be infinite and no further damage calculation is necessary.
If the constant amplitude fatigue limit of the resistance S-N curve corresponds to an
endurance less than 108 cycles, the fatigue resistance curve has to be modified according
to fig. (4)-1. Then the slope m2 of the S-N curve from the constant amplitude fatigue limit
page 107
(5A 106 cycles) up to 108 cycles is assumed to be m2 = 2A m1 - 1 [30] 1.
Fig. (4)-1 Modification of fatigue resistance Wöhler S-N curve for Palmgren-
Miner summation
For fatigue verification it has to be shown that the calculated usable cycles are larger than
the anticipated number of cycles occurring in service of the structure:
1
Although it is accepted that the stresses below the constant amplitude fatigue
limit must be included in cumulative damage calculation relating to welded joints,
there are currently different opinions how this should be achieved. The method
presented here (fig. (4)-1) appears in a number of codes, including Eurocode 3.
However, recent research indicates that it can be unconservative. Here, this questi-
on is partially solved by recommending a Miner sum of ED=0.5 ... 1.0 depending
on shape of the spectrum. Other suggestions recommend that the S-N curve should
be extrapolated further down before the slope change is introduced. For critical
cases or areas of doubt, the user should consult relevant published literature.
page 108
where EDd damage by summation (note restrictions in 4.2 and 4.3)

i index for block number in load spectrum of required design life
ni number of cycles of stress range )Fi,S,d in load spectrum block i
Ni number of cycles at which design stress range )Fi,S,d causes failure in the
modified design fatigue resistance S-N curve.
Nvar number of cycles calculated at variable amplitude load by use of damage
summation EDd
Nconst number of cycles calculated at constant amplitude load of maximum stress
range in spectrum
Nusable number of calculative usable cycles
The order of sequence of the blocks has no effect on the results of this calculation.
In some cases it might be convenient to calculate an equivalent constant amplitude stress

range )FE and to compare it directly to the constant amplitude resistance S-N curve
neglecting the constant amplitude fatigue limit. If all stress ranges of the spectrum are
below the constant amplitude fatigue limit, no )FE needs to be calculated.
For the grid of fatigue resistance classes and an initial slope of m=3 predominantly used in
3.2, the values of the modified characteristic fatigue resistance S-N curves have been
calculated. Stepping down one class corresponds to a division by 1.12. So different levels
of safety (M of S-N curve can be achieved (see 6.4.3).
page 109
Tab. {4}-1: Constants, constant amplitude fatige limit and cut-off limits
Values of modified characteristic fatigue resistance S-N curves for Palmgren-Miner

summation. Initial slope m1=3.0, constant amplitude fatigue limit )FL,k at 5A106
cycles, second slope m2=5.0, cut-off at 108 cycles.
Class Constant C for S-N constant Constant C for S-N cut-off

curve at N<5e6 ampl. fat. curve at N>5e6 limit
cycles, m1=3 limit cycles, m2=5
225 2.278e13 166 6.261e17 91.1
200 1.600e13 147 3.474e17 80.9
180 1.166e13 133 2.052e17 72.8
160 8.192e12 118 1.139e17 64.8
140 5.488e12 103 5.840e16 56.7
125 3.906e12 92.1 3.313e16 50.6
112 2.810e12 82.5 1.913e16 45.3
100 2.000e12 73.7 1.086e16 40.5
90 1.458e12 66.3 6.411e15 36.4
80 1.012e12 58.9 3.558e15 32.4
71 7.158e11 52.3 1.959e15 28.7
63 5.001e11 46.4 1.078e15 25.5
56 3.512e11 41.3 5.980e14 22.7
50 2.500e11 36.8 3.393e14 20.2
45 1.823e11 33.2 2.004e14 18.2
40 1.280e11 29.5 1.112e14 16.2
36 9.331e10 26.5 6.565e13 14.6
32 6.554e10 23.6 3.643e13 13.0
28 4.390e10 20.6 1.869e13 11.3
25 3.125e10 18.4 1.060e13 10.1
22 2.130e10 16.2 5.596e12 8.9
20 1.600e10 14.7 3.474e12 8.1
18 1.166e10 13.3 2.052e12 7.3
16 8.192e9 11.8 1.139e12 6.5
14 5.488e9 10.3 5.840e11 5.7
page 110
4.3.2 Nonlinear Damage Calculation

A nonlinear fracture mechanics damage calculation according to 4.4 is recommended in
cases, where
a) the Miner summation is sensitive to the exact location of the knee point of the
fatigue resistance S-N curve,
b) the spectrum of fatigue actions (loads) varies in service or is changed, and so the
sequence of loads becomes significant or
c) the resistance S-N curve of a pre-damaged component has to be estimated.
Where the parameters for a fracture mechanics fatigue assessment are not known and only
the resistance S-N curve is known, the S-N curve can be used to derive dimensionless
fracture mechanics parameters, which allow a damage calculation [31]. The procedure is
based on the "Paris" power law of crack propagation
where a crack parameter, damage parameter (dimensionless)

N Number of cycles
)K range of stress intensity factor
)Kth threshold value of stress intensity factor range
C0, m material constants
The characteristic stress intensity factor range )KS,k of the fatigue action is calculated with
the stresses of the spectrum )Fi,S,k and the crack parameter a
The characteristic resistance parameters can be derived from the characteristic constant
amplitude fatigue resistance S-N curve: The threshold value corresponds to the fatigue
limit, )Kth,k=)FL,R,k, m equals the slope of the S-N curve, and the constant C0,k can be
calculated from a data point ()FS-N and NS-N) on the S-N curve, preferably from the fatigue
class at 2 A106 cycles
The fatigue verification is executed according to 4.4, using an initial crack parameter ai=1
and a final one af=4 or a large number e.g. af=109. The restrictions on life cycles given in
The actual fatigue class of a pre-damaged component is FATact. = FAT/%a.

4.3 are to be considered.
page 111
4.4 FATIGUE ASSESSMENT BY CRACK PROPA-

GATION CALCULATION
The fatigue action represented by the design spectrum of stress intensity factor ranges
is verified by the material resistance design parameters against crack propagation
using the "Paris" power law
where a crack parameter, damage parameter

N Number of cycles
)K range of stress intensity factor
)Kth threshold value of stress intensity factor range
C0, m material constants
At stress intensity factors which are high compared with the fracture toughness of the
material, Kc, an acceleration of crack propagation will occur. In these cases, the following
extension of the "Paris" power law of crack propagation is recommended. In the absence
of an accurate value of the fracture toughness, a conservative estimate should be made.
where
Kc fracture toughness
R stress ratio
The number of life cycles N is determined by integration starting from an initial crack
parameter ai to a final one af. The calculated number of life cycles N has to be greater or
equal to the required number of cycles.
In general, the integration has to be carried out numerically. The increment for one cycle
is
page 112
It is recommended that a continous spectrum is subdivided to an adequate number of stress

range blocks, e.g. 8 or 10 blocks, and the integration performed blockwise by summing the
increments of a and the number of cycles of the blocks. The entire size of the spectrum in
terms of cycles should be adjusted by multiplying the block cycles by an appropriate factor
in order to ensure at least 20 loops over the whole spectrum in the integration procedure.
4.5 FATIGUE ASSESSMENT BY SERVICE TESTING

4.5.1 General
Components or structures may be tested or verified in respect to fatigue for different
reasons:
a) Existence of a new design with no or not sufficient knowledge or experience of

fatigue behaviour.
b) Verification of a component or structure for a specified survival probability under

a special fatigue action (stress) history.
c) Optimization of design and/or fabrication in respect of weight, safety and economy

after pre-dimensioning. Predimensioning may be done by the use of higher fatigue
resistance data, according to a lower survival probability in comparison with the
resistance data given here. Then the verification is achieved by a subsequent com-
ponent testing.
Fig. (4.5)-1: Example of scatter of test data
A predimesioning leading to the mean values of the resistance data may be done by
page 113
multiplying the resistance values in terms of stress by a factor of 1.47, which is based on a
standard deviation of log cycles of 0.25 and an exponent of m=3.00 .
The verification or assessment depends of the safety strategy considered (see 5.2). Safe life,
fail safe and damage tolerant strategy have to be distinguished.
The fatigue tests should be performed using the data of the fatigue action history (see 3.7),
factored by the partial safety factors (F and (M, i.e., the stress levels of the action history
have to be multiplied by (F @(M for testing.
Tab. {4.5}-1 : Testing approaches

# Testing procedure Approach
1 all specimens of the samples are tested until failure all failed
2 testing is stopped at failure of first specimen of the sample first to fail
3 testing is stopped when p specimens of the n samples fail p to fail
The all failed approach is the normal way of testing at small size samples of which each
specimen represents the same weld details. The statistical analysis uses the data of the
failed specimens disregarding the non-failed ones.
The first to fail approach may be used at a large scale sample with the same weld details
and loading. The test is stopped at the first failure of a specimen.
The n to fail approach is used in similar conditions as the “first to fail” one, when repairs
of crack details can be performed during the test. Each time when a detail fails, the test is
stopped and the failed detail is repaired. Repairs are stopped depending of test conditions.
At the end possibbly all details have failed and thus the “all failed” approach is applied. If
only p specimens out of the n size of the sample failed, the “p to fail” approach is used.
This chapter considers the all failed and first to fail approaches. Other approaches and
details of statistical analysis are considered in appendix 6.4.
The following test result data should be documented according to the selected approach:
C The mean of the log of number of cycles at failure of all n failed samples or details.
C The number of cycles of the first failed detail within n tested details.
C The number of cycles of the first p failed details within n tested details.
The tests should be performed according to well established and appropriate procedures or
standards [32].
For the evaluation of service tests, an estimate of the standard deviation of logN has to be
made, taking into account that the standard deviation varies with the life cycle of the
component to be assessed, see fig. (3.7)-1.
page 114
For the number of test results being n>10, the standard deviation has to be calculated as
given in 3.7-c.
For the number of test results being n<10, or if the procedure of first failure or p failures
in n specimens is used, the standard deviation can be estimated as follows:
C 0.178 for geometrically simple structures at a number of cycles between 104 and
105
C 0.25 for complex structures at cycles up to 106
C ----- no estimate can be given for higher cycles near the endurance limit. Here
special verification procedures are recommended, see ref [32yy]
4.5.2 Acceptance Criteria

The number of design life cycles of the component or structure should be less than the
minimum probable number of the test life cycles.
where
NT number of test life cycles of the test specimens corresponding to the log
mean value or number of cycles of the first test specimen to fail, whichever
is applicable.
F factor dependent of the number of test results available as defined in tables
{4.5}-1 and {4.5}-2. The F-factors refer to a 95% survival probability at a
two sided confidence level of 75% of the mean (see also 6.4)
Nd number of design life cycles, up to which the component or structure may
be used in service
If all components or test specimens are tested to failure, table {4.5}-1 shall be used.
Tab. {4.5}-1: F-factors for failure of all test specimens
Stdv. \ n 2 4 6 8 10
0.178 3.93 2.64 2.45 2.36 2.30

0.200 4.67 2.97 2.73 2.55 2.52
0.250 6.86 3.90 3.52 3.23 3.18
If the tests are carried out until failure of the first test specimen, table {4.5}-2 shall be used
(see also 6.4).
The factor F may be further modified according to safety requirements as given in chapter
5.3. For more details see appendix 6.4.
page 115
Tab. {4.5}-2: F-factors for the first test specimen to fail
Stdv. \ n 2 4 6 8 10
0.178 2.72 2.07 1.83 1.69 1.55

0.200 3.08 2.26 1.98 1.80 1.64
0.250 4.07 2.77 2.34 2.09 1.85
4.5.3 Safe Life Verification
Safe life verification considers each structural element and detail as independent. Each
element has to fulfill the acceptance criteria as defined in 4.5.2.
The partial safety factors (F applied to fatigue actions (loads) and (M applied to fatigue
resistance may be selected from appendix 6.4.3.
4.5.4 Fail Safe Verification
Fatigue life verification of fail safe structures depends largely on the design and operation
parameters of a structure. The effectivness of statically over-determined (hyperstatic)
behaviour or redundancy of structural components, the possibility of detection of failures
in individual structural parts and the possibility of repair determine the level of safety
required in the individual structural parts. So, no general recommendation can be given.
It is recommended that the factor F given in 4.5.2 is used as a general guidance and to
establish agreement.
4.5.5 Damage Tolerant Verification
The verification is based on crack growth measurements, starting from a crack size, which
can be detected in inspection up to a critical crack size, at which the limit state of critical
safety against brittle or plastic fracture or other modes of failure of the remaining sectional
area is attained.
The criteria for factoring the observed life cycles for the test depend of the application. It
is recommended to establish agreement on the factor F.
page 116
5 SAFETY CONSIDERATIONS
A component has to be designed for an adequate survival probability. The required survival
probability is dependent on the
a) uncertainties and scatter in the fatigue assessment data,

b) safety strategy and
c) consequences of failure.
The uncertainties of fatigue assessment data may arise from fatigue actions, such as
1. determination of loads and load history,

2. determination of stresses or stress intensity factors from the model used for
analysis, and
3. dynamic response problems.
These uncertainties are covered by an appropriate partial safety factor for the fatigue
actions (F, which is not considered here.
Uncertainties of fatigue assessment data arising from fatigue resistance and damage
calculation are:
4. scatter of fatigue resistance data,

5. scatter of verification results of damage calculations.
The sources of uncertainty numbered 4. and 5. are considered here. For normal ap-
plications, they are already covered in the fatigue resistance data given here. For special
applications, the data may be modified by the selection of an adequate partial safety factor
(M.
5.2 FATIGUE DESIGN STRATEGIES

Different ways of operation in service require different fatigue design strategies. The
definition of a fatigue strategy refers predominantly to the method of fatigue analysis,
inspection and monitoring in service.
5.2.1 Infinite Life Design

This strategy is based on keeping all fatigue actions under the resistance fatigue limit or
threshold value. No regular monitoring in service is specified. So a high survival probability
has to be provided. For fatigue actions which are almost uniform and act at very high
page 117
cycles this strategy may be adequate.
5.2.2 Safe Life Design
This design strategy is based on the assumption that initially the welded joint is free of
imperfections. No regular monitoring in service is specified, so a high survival probability
has to be provided.
5.2.3 Fail Safe Design
This design strategy is based on statically over-determined (hyperstatic) or redundant

structures. No regular monitoring in service is provided. In case of a fatigue failure,
redistribution of forces provides an emergency life, so that the failure can be detected and
repaired. The welded joints can be designed for a normal survival probability.
5.2.4 Damage Tolerant Design

This design strategy is based on the assumption of the presence of cracks as large as the
detection level of the non-destuctive testing method applied. Fracture mechanics is used to
calculate the life cycles until failure. From the number of life cycles, regular inspection
intervals are derived. A normal probability of survival is adequate.
5.3 PARTIAL SAFETY FACTORS

The required partial safety factor (M depends largely on circumstances such as
a) fatigue design strategy

b) consequences of failure
c) practical experience in fields of application.
Thus, no general recommendation can be given. In most cases for normal fabrication
quality and regular inspections in service, (M=1 might be adequate.
The safety factors are given in terms of stress. If safety factors are needed in terms of
cycles, 'M may be calculated using the slope m of the resistance S-N curve
It should be recognized that the slope m of the S-N curve varies with the number of cycles,
see fig. (3.7)-1. An example of a possible table of partial safety factors is given in appendix
6.4.
page 118
5.4 QUALITY ASSURANCE

Weld quality is assured preferably by the application of an ISO 9000 quality management
system or a comparable system.
The weld quality should be equal to quality class B according to ISO 5817. However,
some exceptions may be allowed in the tables given in chapter 3.2.
Besides regulations and quality codes, the general standards of good workmanship have to
be maintained.
5.5 REPAIR OF COMPONENTS

The most common cause of damage of welded structures and components is fatigue.
Before any start of repair actions, it is vitally important to establish the reason for the
damage.This will influence the decisions to be made about the need for repair and for the
repair methods. Possible reasons for fatigue damage may be:
C inadequate load assumption with regard to stress range, number of cycles and shape
of load spectrum
C inadequate stress analysis
C inadequate structural design, especially of weld details
C inadequate material e.g. regarding toughness and weldability
C inadequate workmanship (parts missing or not properly positioned, unsatisfactory
application of thermal cutting processes, excessive weld imperfections as e.g. poor
penetration, severe undercut, severe misalignment, unauthorized welding e.g. of
fabrication aids)
C resonant or non-resonant forced vibrations or dynamic response not expected or
not considered in design
C environmental influences enhancing fatigue such as corrosion or elevated
temperature
C faulty operation, e.g. overload or fretting
C accident, e.g. collision
In most cases of damage, design, loads and imperfections are the governing parameters of
the failure, material properties are often secondary.
The actions to be taken should be based on the results of the investigations. Possible
actions are:
C no repair
C delayed repair
C immediate repair
C more frequent or contonuous crack monitoring, in-service inspection or vibration
page 119
monitoring
C change in operation conditions
A large variety of repair methods exist. They may generally include the following aspects:
C removal of crack
C modification of detail design
C modification of service loads
C selection of adequate material and repair welding procedure
C weld toe improvements techniques
C quality control of the repair weld
page 120
6 APPENDICES
The appendices are intended to give special guidances, background information and
additional explanations. They are not normative.
6.1 LOAD CYCLE COUNTING

6.1.1 Transition Matrix
To establish a transition matrix, first the number of different stess levels or classes has to
be defined. 32 stress levels are sufficient. Then the numbers (occurrence) of transitions
(reversals) from one extreme value (peak or through) to another are counted and
summarized in the matrix. A number in the matrix element ai, j indicates the number of
transitions from a stress belonging to class i to a stress belonging to class j.
Fig. (6.1)-1 Principle of the transition matrix
The data for the transition matrix can be obtained by measurement or by time simulation
computations. A time signal for fatigue tests or crack propagation simulations or
cumulative frequency diagrams (stress spectra) for damage calculations can be generated
from the transition matrix by a Markov random draw.
6.1.2 Rainflow or Reservoir Counting Method
The algorithm of rainflow counting method is well explained by using the analogy of the
flow of water on a pagoda roof. The stress signal, looked at vertically, is regarded as the
pagoda roof. A cycle is obtained, when a contour is closed by the drop of the flow from a
peak to a slope of the roof [26 and 27]. The range is then equal to the difference between
the extreme values of the contour. Later the smaller included cycles can be determined the
same way. The non closed contour from the extreme of the entire signal leads to a half
page 121
cycle. Reservoir counting is similar.
Fig. (6.1)-2 Illustration of rainflow counting
6.2 FRACTURE MECHANICS

6.2.1 Rapid Calculation of Stress Intensity Factors
A simplified method may be used to determine Mk-factors [19]. Here, the Mk-factors are
derived from the non-linear stress peak distribution Fnlp(x) along the anticipated crack path
x assuming no crack being present. Hence, the function of the stress concentration factor
kt,nlp(x) can be calculated. The integration for a certain crack length a yields:
For different crack lengths a, a function Mk(a) can be established, which is preferably
presented in the form:
6.2.2 Dimensions of Cracks
Tab. {6.2}-1: Dimensions for assessment of crack-like imperfections (example)
Idealizations and dimensions of crack-like imperfection for fracture mechanics

assessment procedure (t = wall thickness).
page 122
6.2.3 Interaction of Cracks

Adjacent cracks may interact and behave like a single large one. The interaction between
adjacent cracks should be checked according to an interaction criterion.
There are different interaction criteria, and in consequence no strict recommendation can
be given. It is recommended to proceed according to an accepted code, e.g. [24].
page 123
6.2.4 Formulae for Stress Intensity Factors

Stress intensity factor formulae may be taken from literature, see references [13] to [20].
For the majority of cases, the formulae given below are sufficient.
Tab. {6.2}-2: Stress intensity factors at welds

Surface cracks under membrane stress
The formula for the stress in-
tensity factor K1 is valid for
a/c < 1,
for more details see ref. [14]
K1 = F /(BAa / Q) A Fs
Q = 1 + 1.464 (a/c)1.65
Fs = [M1 + M2A(a/t)2 + M3A(a/t)4]AgAfAfw
M1 = 1.13 - 0.09 (a/c)

M2 = -0.54 + 0.89 / (0.2 + a/c)
M3 = 0.5 - 1 / (0.65 + a/c) + 14 (1 - a/c)24
fw = [sec(BAc /(a/t) /(2Ab))] 1/2
g and f are dependent to direction
"a"-direction: g = 1 f = 1
"c"-direction: g = 1 + [0.1 + 0.35 (a/t)2] f = /(a/c)
page 124
Embedded cracks under membrane stress

a/c < 1,
for more details see ref. [14]
K1, Q, Fs, fw as given before for surface cracks, but:
M1 = 1
M2 = 0.05 / (0.11 + (a/c)3/2)
M3 = 0.29 / (0.23 + (a/c)3/2)
"c"-direction: g = 1 - (a/t)4 / (1 + 4a/c) f = /(a/c)
Surface cracks under shell bending and membrane stress

a/c < 1,
for more details see ref. [14].
K1 = (Fmem + HAFben) /(BAa / Q) A Fs
Q = 1 + 1.464 (a/c)1.65
Fs = [M1 + M2A(a/t)2 + M3A(a/t)4]AgAfAfw
M1 = 1.13 - 0.09 (a/c)

M2 = -0.54 - 0.89 / (0.2 + a/c)
M3 = 0.5 - 1 / (0.65 + a/c) + 14 (1 - a/c)24
fw = [sec(BAc /(a/t) /(2Ab))] 1/2
The function H is given by the formulae:
"a"-direction: H = 1 + G1(a/t) + G2(a/t)2
where G1 = -1.22 -0.12A(a/c)

G2 = 0.55 - 1.05A(a/c)0.75 + 0.47(a/c)1.5
"c"-direction: H = 1 - 0.34 (a/t) - 0.11 (a/c) (a/t)
Surface crack in cylinder under internal pressure
page 125

a/c < 1,
for more details see ref. [15],
where D is the diameter in mm
and P is the internal pressure
in N/mm2.
K1 = F /(BAa / Q) A Fs
S = pADinner / (2t)
Q = 1 + 1.464 (a/c)1.65
Fs = 0.97A[M1 + M2A(a/t)2 + M3A(a/t)4]ACAgAfAfw
M1 = 1.13 - 0.09 (a/c)

M2 = -0.54 - 0.89 / (0.2 + a/c)
M3 = 0.5 - 1 / (0.65 + a/c) + 14 (1 - a/c)24
fw = [sec(BAc /(a/t) /(2Ab))] 1/2
C = [(Dout2 + Din2) / (Dout2 - Din2) + 1 - 0.5A/ a/t ]A2t/Din
Through the wall cracks in curved shells under internal pressure

In sphere and longitudinal
cracks in cylinder loaded by
internal pressure. Mk covers
increase of stress
concentration factor due to
bulging effect of shell. For
details see ref. [15,18].
K = Fmem A /(BAa) A Mk
where Mk = 1.0 for x < 0.8
and Mk = /(0.95 + 0.65Ax - 0.035Ax1.6) for x > 0.8

and x < 50
with x = a / /(rAt)
a half distance between crack tips of through the wall crack

r radius of curvature perpenticular to the crack plane
t wall thickness
page 126
Root gap crack in a fillet welded cruciform joint

tensity factor K is valid for
H/t from 0.2 to 1.2 and for a/w
from 0.0 to 0.7. For more de-
tails see ref. [17].
FA(A + A A a/w) A / (BAa A sec(BAa/2w))

1 2
K = _______________________________________________________________________________
1 + 2AH/t
where w = H + t/2
F = nominal stress range in the longitudinal plates
and with x = H/t
A1 = 0.528 + 3.287Ax - 4.361Ax2 + 3.696Ax3 - 1.875Ax4 + 0.415Ax5

A2 = 0.218 + 2.717Ax - 10.171Ax2 + 13.122Ax3 - 7.755Ax4 + 1.783Ax5
For a variety of welded joints parametric formulae of the Mk functions have been
established and published [18,19]. For the majority of cases, the formulae given below are
sufficient [20].
Tab. {6.2}-3:
Weld local geometry correction for crack at weld toe
Applicable for transverse
full penetrating or non-
loadcarrying welds, e.g.
butt weld, transverse
attachment,cruciform
joint K-butt weld. For
more details see ref.
[20].
Stress intensity magnification factor Mk > 1 for membrane stress:

for l/t # 2:
Mk = 0.51A(l/t)0.27A(a/t)-0.31 , for (a/t) # 0.05A(l/t)0.55

Mk = 0.83A(a/t)-0.15(l/t)^0.46 , for (a/t) > 0.05A(l/t)0.55
for l/t > 2:
Mk = 0.615A(a/t)-0.31 , for (a/t) # 0.073

Mk = 0.83A(a/t)-0.2 , for (a/t) > 0.073
Stress intensity magnification factor Mk > 1 for bending stress:
for l/t # 1:
Mk = 0.45A(l/t)0.21A(a/t)-0.31 , for (a/t) # 0.03A(l/t)0.55
Mk = 0.68A(a/t)-0.19(l/t)^0.21 , for (a/t) > 0.03A(l/t)0.55
for l/t > 1:

Mk = 0.45A(a/t)-0.31 , for (a/t) # 0.03
Mk = 0.68A(a/t)-0.19 , for (a/t) > 0.03
page 127
A systematic set of formulae was also developed in [xx] using the procedure outlined in
chapter 6.2.1. The formulae are valid within the given dimensional validity ranges.
Tab. {6.2.4}: Formulae for Mk values for different welded joints

Transverse non-loadcarrying attachment
Dim. min max
H/T 0.2 1
W/T 0.2 1
2 15° 60°
A/T 0.175 0.72
t/T 0.125 2 (4)
(1)
Cruciform joint K-butt weld

Dim. min max
H/T 0.2 1
W/T 0.2 1
2 15° 60°
A/T 0.175 1.3
t/T 0.5 20
(2)
page 128
Cruciform joint fillet welds

Dim. min max
H/T 0.2 1
W/T 0.2 1
2 15° 60°
A/T 0.175 0.8
t/T 0.5 10
If 0.2 < H/T < 0.5 and 0.2 < W/T < 0.5 and a/T < 0.07 then:
(3)
If 0.2 < H/T < 0.5 and 0.2 < W/T < 0.5 and a/T > 0.07 then:
(4)
If 0.5 < H/T < 1.5 or 0.5 < W/T < 1.5 then:
(5)
page 129
Lap joint
Dim. min max
H/T 0.25 1
W/T 0.25 2
U/T 0 1.5
2 15° 70°
A/T 0.175 0.7
t/T 0.3 1
(6)
Longitudinal non-loadcarrying attachment

Dim. min max
L/T 5 40
B/T 2.5 40
2/45° 0.670 1.33
t/T 0.25 2
A = 0.7 A t
(7)
page 130
6.3 FORMULAE FOR MISALIGNMENT

Tab. {6.3}-1: Formulae for assessment of misalignment
# TYPE OF MISALIGNMENT
1 Axial misalignment between flat plates
8 is dependent on restraint, 8=6 for unrestrained joints.

For remotely loaded joints assume l1=l2.
2 Axial misalignment between flat plates of differing thickness
Relates to remotely loaded unrestraint joints.

The use of n=1.5 is supported by tests.
3 Axial misalignment at joints in cylindrical shells with thickness change
n=1.5 in circumferential joints and joints in spheres.

n=0.6 for longitudinal joints.
page 131
4 Angular misalignment between flat plates
Assuming fixed ends:
assuming pinned ends:
The tanh correction allows for reduction of angular misalignement due to the
straightening of the joint under tensile loading. It is always #1 and it is
conservative to ignore it.
5 Angular misalignment at longitudinal joints in cylindrical shells

Assuming fixed ends:
assuming pinned ends:
d is the deviation from the idealized geometry
page 132
6 Ovality in pressurized cylindrical pipes and shells
7 Axial misalignment of cruciform joints (toe cracks)
8 is dependent on restraint
8 varies from 8=3 (fully restrained) to 8=6 (unrestraint). For unrestrained

remotely loaded joints assume: l1=l2 and 8=6
8 Angular misalignment of cruciform joints (toe cracks)
8 is dependent on restraint
If the inplane displacement of the transverse plate is restricted, 8 varies from

8=0.02 to 8=0.04. If not, 8 varies from 8=3 to 8=6.
page 133
9 Axial misalignment in fillet welded cruciform joints (root cracks)
km refers to the stress range in weld throat.
page 134
6.4 STATISTICAL CONSIDERATIONS ON SAFETY

6.4.1 Statistical Evaluation of Fatigue Test Data
The different methods as described in 3.7 consider different statistical effects, when
evaluating a set of fatigue data. Ideally, all effects have to be considered, e.g.
a) Variance of data
b) Probability distribution of the mean value by its confidence interval
c) Probability distribution of the variance by its confidence interval
c) Difference of the distribution of the whole set of data (population) and the
distribution of the sample (Gaussian versus t-distribution)
d) Deviation from the assumed Gaussian distribution which can be evaluated
by a likelihood or a P2 test
For design, a safety margin is considered, which is applied to the mean values. The values
used for design are the so called characteristic vakues (index k).
These charactreistic values are, in principle, values at a "=95% survival probability (5%
probability of failure) associated to a two sided confidence interval of 75% of the mean xm
and of the standard deviation Stdv, i.e. $=75% (12.5% probability of being above or
below the extreme value of the confidence interval):
The factor ki considers the effects a) to d) and corresponds to:
C the minimun value of the mean confidence interval

C the maximum value of the variance confidence interval
Taking into account that the probability distribution of the mean corresponds to a Student
law (t-distribution) and the probability distribution of the variance corresponds to a Chi-
square law (P2), the general formula for ki is given by:
where t value of the two sided t-distribution (Student’s law) for p=$=0.75, or of
the one sided t-distribution for a probability of p=(1+$)/2=0.875 at n-1
degrees of freedom
n number of data (test specimens of details)
N distribution function of the Gaussian normal distribution probability of
exceedence of "=95% (superscript -1 indicates inverse function)
P2 Chi-sqare for a probability of (1+$)/2=0.875 at n-1 degrees of freedom
page 135
If the variance is fixed from other tests or standard values, no confidence interval has to be
considered and so the factor is given by:
Tab. {6.4}-1: k-values for the different methods

n t P2 k1 k2
2 2.51 2.80 11.61 3.41
3 1.61 0.27 5.41 2.57
4 1.44 0.69 4.15 2.36
5 1.36 1.21 3.60 2.25
10 1.24 4.47 2.73 2.04
15 1.21 8.21 2.46 1.96
20 1.20 12.17 2.32 1.91
25 1.19 16.26 2.24 1.88
30 1.18 20.45 2.17 1.86
40 1.18 29.07 2.09 1.83
50 1.17 37.84 2.04 1.81
100 1.16 83.02 1.91 1.76
6.4.2 Statistical Evaluation at Component Testing
Testing all test specimens to failure
When all specimens are tested to failure, the procedure is to estimate the mean log NT of
the S-N curve and the associated standard deviation.
Starting from the formula in 4.5.2, there is
which defines the safety factor F by:
Taking the acceptance criterion from chapter 3.7 xm - k Stdv > xk the factor F can be
received:
With the formula for k the different values of F can be calculated, depending on number
page 136
of test specimens n and on the assumed standard deviation Stdv of the test specimens in
terms of logN.
Testing all test specimens simultaneously until first failure
When all test specimens are tested simultaneously until the first to fail, only one value of
log NT is obtained and no standard deviation can be derived from test results.
Starting from the formula in 4.5.2, there is
which defines the safety factor F by:
When considering statistical evaluation, account must be taken of additional effects as

illustrated in fig. (6.4)-1:
1) Distribution of the 1/n-th extreme

value
2) Distribution of the sample between
1/n-th extreme and mean
3) Safety margin for the characteristic
value
where NT first failure

xm mean of the sample
Nk charactristic value
Nd design value Fig. (6.4)-1 Distribution of action and re-
sistance
log NT is considered as the probable
maximum (safe side) of the distribution of the minimum value of the log N distribution. The
mean sample xm is therefore given by:
with Stdv standard deviation of the sample

" from table of variance order statistics
ka, kb from table of expected values of normal order statistics
Taking the acceptance criterion from chapter 3.7, xm - k1 Stdv > xk , the factor F can be
received:
The different values of F can be calculated, depending on number of test specimens n and
on the assumed standard deviation Stdv of the test specimens in terms of log N.
page 137
Tab. {6.4}-3: Values k for testing until first failure

n 2 4 6 8 10
k 2.44 1.77 1.48 1.28 1.07
For more details see ref. [35, 36].
Testing all specimens simultaneously until p failures amongst n specimens
Values of k may be taken from the relevant literature or from IIW doc. XIII-1822-2000
(under development).
6.4.3 Statistical Considerations for Partial Safety Factors
No general recommendations on partial safety factors are given. For special fields of
application, tables of safety factors on load actions (F and on fatigue resistance (M may be
established. Table {6.4}-4 shows a possible example for (M which may be adjusted
according to the special requirements of the individual application.
Tab. {6.4}-4: Possible example for partial safety factors (M for fatigue resistance
Partial safety factor (M 6 Fail safe and damage Safe life and infinite life
Consequence of failure tolerant strategy strategy
Loss of secondary 1.0 1.15

structural parts
Loss of the entire structure 1.15 1.30
Loss of human life 1.30 1.40
page 138
7 REFERENCES
Discussion: More references needed??? Please check and give me!
General:
[1] ISO 2394

General principles on reliability for structures.
Second edition 1986-10-14
[2] Niemi E.
Recommendations concerning stress determination for fatigue analysis of welded
components.
IIW doc. XIII-1458-92/XV-797-92
[3] Gurney T.R.

Fatigue of Welded Structures.
Cambridge University Press, UK, 1978
[4] Maddox, S.J.

Fatigue Strength of Welded Structures.
Abington Publishing, Abington UK, 1991
[5] Radaj D.
Design and analysis of fatigue resistent welded structures
Abington Publishing, Abington Cambridge, U.K. 1990
[6] Hobbacher A. et al.

Design recommendations for cyclic loaded welded steel structures
IIW doc. XIII-998-81/XV-494-81; Welding in the World, 20(1982), pp. 153-165
Structural hot spot stress procedure:
[7] Huther M. and Henry J.

Recommendations for hot spot stress definition in welded joints.
IIW doc. XIII-1416-91
[8] Huther M, Parmentier G. and Henry J.

Hot spot stress in cyclic fatigue for linear welded joints.
IIW doc. XIII-1466-92/XV-796-92
[x1] Niemi E.
Strucural Stress Approach to Fatigue Analysis of Welded Components - Designer’s
Guide -. IIW doc. XIII-1819-00 / XV-1090-01
page 139
x2 Recommended fatigue design procedure for welded hollow section joints.

IIW doc. XIII-1772-99 / XV-1021-99
Effective notch stress procedure:
[9] Petershagen H.
A comparison of approaches to the fatigue strength assessment of welded
components
IIW document XIII-1208-86, 1986
[10] Petershagen H.
Experiences with the notch stress concept according to Radaj (transl.)
15. Vortragsveranstaltung des DVM Arbeitskreises Betriebsfestigkeit, Ingolstadt
18.-19.10.1989
[11] Olivier R., Köttgen V.B., Seeger T.

Welded connection I: Fatigue assessment of welded connections based on local
stresses (transl.)
Forschungskuratorium Maschinenbau, Bericht No. 143, Frankfurt 1989 (143 pages)
[12] Köttgen V.B., Olivier R., Seeger T.

Fatigue analysis of welded connections based on local stresses
IIW document XIII-1408-91, 1991
Fracture mechanics:
[13] Murakami Y.
Stress Intensity Factors Handbook
Pergamon Press, Oxford U.K. 1987
[14] Newman J.C. and Raju I.S.

Stress intensity factor equations for cracks in three-dimensional finite bodies.
ASTM STP 791 1983, pp. I-238 - I-265.

Stress intensity factors for internal surface cracks in cylindrical pressure vessels.
Journal of Pressure Vessel Technology, 102 (1980), pp. 342-346.

An empirical stress intensity factor equation for the surface crack.
Engineering Fracture Mechanics, vol 15. 1981, No 1-2, pp. 185-192.
[17] Frank K.H. and Fisher J.W.

Fatigue strength of fillet welded cruciform joints.
J. of the Structural Div., Proc. of the ASCE, vol 105 (1979) pp. 1727-1740
page 140
[18] Folias E.S.

Axial crack in pressurized cylindrical shell.
Int. J. of Fracture Mechanics, vol 1 (1965) No. 2, pp 104
[19] Hobbacher A.
Stress intensity factors of welded joints.
Engineering Fracture Mechanics, vol 46 (1993), no 2, pp. 173-182, et vol 49
(1994), no 2, p. 323.
[20] Maddox S.J., Lechocki J.P. and Andrews R.M.

Fatigue Analysis for the Revision of BS:PD 6493:1980
Report 3873/1/86, The Welding Institute, Cambridge UK
Fatigue strength modifications:
[21] Ørjasæter, O.
Effect of plate thickness on fatigue of welded components.
IIW doc. XIII-1582-95 / XV-890-95
Haagensen P.J. and Maddox S.J.: IIW Recommendations for Weld Toe
Improvement by Grinding, TIG Dressing and Hammer Peening for Steel and
Aluminium Structures.
IIW doc. XIII-1815-00 (2000).
Weld imperfections:
[22] IIW guidance on assessment of the fitness for purpose of welded structures.
IIW doc. SST-1157-90
[23] Hobbacher A. et al.

Recommendations for assessment of weld imperfections in respect of fatigue.
IIW doc. XIII-1266-88/XV-659-88
[24] Guidance on some methods for the derivation of acceptance levels for defects in
fusion welded joints.
British Standard Published Document 6493:1991
[25] Ogle M.H.

Weld quality specifications for steel and aluminium structures.
Welding in the World, Vol. 29(1991), No, 11/12, pp. 341-362
Stress spectrum:
[26] Endo T. et al.

Fatigue of metals subjected to varying stress - prediction of fatigue lives (transl.)
page 141
Kyushu District Meeting of the JSME, Nov. 1967.

also: Rain flow method - the proposal and the applications.
Memoir Kyushu Institut of Technical Engineering, 1974.
[27] Standard Practice for Cycle Counting in Fatigue Analysis.

ASTM E 1049-85
Damage calculation:
[28] Palmgren, A.
On life duration of ball bearings (transl.).
VDI-Z. vol. 68(1924), pp 339-341
[29] Miner, A.M.

Cumulative damage in fatigue.
J. Appl. Mech. September 1945. pp 151-164.
[30] Haibach E.
Modified linear damage accumulation hypothesis considering the decline of the
fatigue limit due to progressive damage (transl.)
Laboratorium für Betriebsfestigkeit, Darmstadt, Germany, Techn. Mitt. TM 50/70
(1970)
[31] Hobbacher A.
Cumulative fatigue by fracture mechanics.
Trans. ASME Series E, J. Appl. Mech. 44(1977), pp. 769-771
Fatigue testing:
[32] Lieurade H.P.

Fatigue Testing of Welded Joints
IIW doc. XIII-1516-93 (ISO porposal)
Quality and safety considerations:
[33] ISO 6520:1982 (EN 26520:1982)

Weld irregularities
[34] ISO 5817:1992 (EN 25817:1992)

Quality groups of welds
[35] Huther M.
Uncertainties, Confidence Intervals and Design Criteria
page 142
IIW dec. XIII-1371-90
[36] Maddox S.J.

Statistical Analysis of Fatigue Data Obtained from Specimens Containing many
Welds
IIW doc. JWG-XIII-XV-122-94
[37] Marquis G. and Mikkola T.

Analysis of welded structures with failed and non-failed welds, based on maximum
likelyhood
IIW document XIII-1822-00
[38] Petershagen H.
IIW Recommendations on the Repair of Fatigue-Loaded Welded Structures.
IIW doc. XIII-1632-96
page 143

IIWrec 2004 02

Uploaded by

Copyright:

Available Formats

IIWrec 2004 02

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IIWrec 2004 02

Uploaded by

Copyright:

Available Formats

INTERNATIONAL INSTITUE INSTITUT INTERNATIONAL

GERMAN DELEGATION DÉLÉGATION ALLEMANDE

Prof. Dr. A. Hobbacher

IIW document XIII-1965-03 / XV-1127-03

RECOMMENDATIONS FOR FATIGUE DESIGN

Update February 2004

Prof. Dr. A. Hobbacher, Univ. Appl. Sc. Wilhelmshaven, Germany, as Chairman

Prof. Dr. H. Fricke, Univ. Hamburg-Harburg, Germany

Prof. Dr. A. Hobbacher

2 FATIGUE ACTIONS (LOADING) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.8.5.2 Simplified Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

4 FATIGUE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

5 SAFETY CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

1.2 SCOPE AND LIMITATIONS

Design value Characteristic value factored by a partial safety factor.

Fatigue Detoriation of a component caused by crack initiation and/or by

Fatigue action Load effect causing fatigue.

Fatigue life Number of stress cycles of a particular magnitude required to

Fatigue limit Fatigue strength under constant amplitude loading corre-

Fatigue resistance Structural detail's resistance against fatigue actions in terms of

Fatigue strength Magnitude of stress range leading to a particular fatigue life.

Local nominal stress Nominal stress including macro-geometric effects, concentrated

Macro-geometric effect A stress raising effect due to macro-geometry in the vicinity of

Miner sum Summation of individual fatigue damage ratios caused by each

Misalignment Axial and angular misalignments caused either by detail design

Modified nominal stress See 'Local nominal stress'.

Nominal stress A stress in a component, resolved using general theories, e.g.

Notch stress concentration

Paris' law An experimentally determined relation between crack growth

Rainflow counting A standardized procedure for stress range counting.

Range counting A procedure of determining various stress cycles and their

S-N curve Graphical presentation of the dependence of fatigue life N on

fatigue strength S ()FR or )JR), also known as Wöhler curve.

Stress cycle A part of a stress history containing a stress maximum and a

Stress history A time based presentation of a fluctuating stress, defined by

Stress range exceedances A tabular or graphical presentation of the cumulative frequency

Stress range occurrences A tabular or graphical presentation of stress ranges, usually

Stress ratio Ratio of minimum to maximum algebraic value of the stress in

Structural discontinuity A geometric discontinuity due to the type of welded joint,

Structural stress A stress in a component, resolved taking into account the

and shell bending stress components. Also referred to as

1.5 BASIC PRINCIPLES

1.6 NECESSITY FOR FATIGUE ASSESSMENT

In the following cases, detailed fatigue assessment is not required:

a) The highest nominal design stress range satisfies

(M should be taken from an applicable design code.

This paragraph is not applicable to tubular joints.

for steel )Kth=2.0 MPa/m

1.7 APPLICATION OF THE DOCUMENT

Tab. {1}-3: General guidance for the application of the document

(1) Does joint determine look up go to (6)

(2) Is Structural determine look up re-

(3) Is effective determine look up re-

(4) Are cracks or determine look up

(5) Test entire go to (8)

Modifications and Assessment Procedures

(7) calc. design perform crack

2 FATIGUE ACTIONS (LOADING)

2.1 BASIC PRINCIPLES