Topics in Ship Structures

01 Fatigue Strength

Reference : 선박해양구조역학 by 고대은 장범선

Fatigue Strength of Welded Stricutres
DNV 30.7 and DNV RP-C 203
2017.09
Naval Architecture and Ocean Engineering
Jang, Beom Seon
OPen INteractive Structural Lab
Introduction
Stress Concentration
2

 Different S-N curves for different stress concentrations.

 Stress concentration due to discontinuity or
change of shape.
 Under static loading, a stress concentration
in a ductile material has no effect on the
strength.

Effect of stress concentration Stress concentration at increase

on fatigue strength of section
2 OPen INteractive Structural Lab
Introduction
Stress concentration caused by weld shape
3

 Stress concentration level depending on geometric transition,

smooth transition vs. abrupt change
 Concentration at the ‘toes’ of weld is the most likely site.
 Weld surface irregularities, like weld ripples, and lumps.

Stress concentration at toe of weld bead Comparison of butt and fillet weld shape

Fatigue cracking from weld ripples, and stop/start positions

3 OPen INteractive Structural Lab
Introduction
4
Alternate Stress and Mean Stress
 Stress range    max   min
 max   min
 Stress amplitude  a 
2
 max   min
 Mean stress m 
2

 Stress Ratio
 min
R
 max
Fully reversed
 For Example
 Fully reversed
 min   max , R  1,   2 max ,  mean  0
 Zero to max
 min  0, R  0,    max ,  mean   max / 2 Zero to max
 Zero to min
 max  0, R  ,    min ,  mean   min / 2
Zero Structural
OPen INteractive to min Lab
Introduction
5
Definition of Three Kinds of Stresses
 Nominal stress
 A stress to be calculated using general theories such as beam theory.
 It is difficult to consider stress concentration caused by geometric discontinuity.
 Hotspot stress (Geometric stress)
 A stress on the surface at hot spot. Not real stress.
 The effect of structural details are taken into account but, the effect of weld bead
is not included.
 To be calculated from shell FE model where weld bead is not included.
 Notch stress
 Total stress taking into account the stress concentration caused by the weld bead
and geometric discontinuity.

OPen INteractive Structural Lab

Introduction
6
Fatigue Fracture
 Three stages of fatigue failure
1) Crack Initiation :
 plastic deformation is accumulated at the toe of weld bead or notch
 Fatigue crack initiates from crack-like flaws and crack growth phase
accounts for 90% of entire fatigue life for welded joint.
2) Crack growth :
 Crack grows perpendicular to principal stress direction.
 Crack growth at crack tip depends on the stress level of crack tip. Crack
always grows under cyclic tensile load.
 Under compressive load, two faces contact and crack tip is closed
3) Final failure : brittle facture, ductile fracture, plastic collapse

Notch crack
propagation under
tensile load

OPen INteractive Structural Lab

Introduction
S-N curve
7

 S-N Curve : Relationship curve between stress range () and number of
stress cycles (N) in log-log scale graph.
 Mean S-N curve
1 1
log    log N  log a
m m
 Design S-N curve
 Survival limit of 97.7% and
failure probability of 2.3%
1 1
log    log N  (log a  2 log s )
m m
1 1 a
  log N  log a, a  2
m m s

s : standard deviation of log N

1
 : slope of S-N curve, m=3,4,5
m

S-N curve

7 OPen INteractive Structural Lab

Introduction
S-N curve
8

 S-N curve depends on the level of stress concentration

 S-N curve is inversely proportional to cubic of stress range.

1 1
log    log N  log a
m m

N  a   m

Example of S-N Curve for three types

8 OPen INteractive Structural Lab

Introduction
Fatigue Test
9

 Cyclic load of 5-15Hz under mean tensile stress to avoid buckling

failure if possible.
 8-24 days for 107 cycles.
 Variations due different weld bead shape.

9 OPen INteractive Structural Lab

Fatigue Failure at Weld Bead
Stress concentration caused by discontinuity
10

 Stress concentration at the edge of a hole

≈ that of Weld toe, but, geometry of weld
toe is different.
 Features of weld toe
• Undercutting or abrupt & convex profile,
• very small crack-like discontinuities, termed
“ intrusions”.
• Variations along welding length 104
→ Kt calculation or experimental technique
are meaningless. Comparison of fatigue strengths

Unfavourable characteristics of
Fatigue cracking from pre-existing
weld profiles
crack-like flaw
10 OPen INteractive Structural Lab
Fatigue Failure at Weld Bead
Stress concentration caused by discontinuity
11

 Welded details : Smaller portion of fatigue crack initiation time than

unwelded details. Crack propagation time is dominant.
 It is essential that welds are made in accordance with strict procedure.
 Weld root fatigue : more severe than weld toe since it is hard to be detected.

Difference between plate and Fatigue cracking from the root

welded joint

11 OPen INteractive Structural Lab

Factors affecting on Fatigue of Welded Joints
MATERIAL PROPERTY
12

 The fatigue strength of unwelded components increases with

material strength.
 However, this is not the case with welded material. ← crack
propagation of crack-life flaws is dominant and the rate of crack
growth is little affected by material strength
► Use of high-tensile steel (‘70~’80)→ high applied stress → fatigue
crack → high repair cost → cap on the use
→ the same S-N curve as that for mild steel

 Brittle material (fracture failure

with small crack) : little affected.
← fatigue crack growth rate
increases exponentially and the
crack size at final failure doesn't
not have a large influence on
fatigue life.

Effect of steel strength on

12 fatigue strength
OPen INteractive Structural Lab
Factors affecting on Fatigue of Welded Joints
Weld quality
13

 Welding flaws like porosity(다공성), slag, intrusions, lack of fusion,

or incomplete weld root penetration reduces fatigue strength.
 Alternative sites for fatigue crack initiation.
 Misalignment like axial eccentricity or angular distortion subjected to
perpendicular load. → local secondary bending → stress increases.
 Strict welding procedure

 Fitness-for-purpose : estimation of
effect of welding flaws and welding
imperfection
→ establishment of acceptance
limit

(a) Axial eccentricity

(b) Angular distortion

13 OPen INteractive Structural Lab
Factors affecting on Fatigue of Welded Joints
Residual Stress
14

 Two systems of stress in welded stresses

1)Reaction stress : from assembly conditions, affecting the various
members as whole.
2) Residual stresses : from weld heating and cooling, affecting localized
area, along welding line.
 During cooling stage the longitudinal shrinkage of the weld metal is
resisted. → high residual tensile stress acting along the welding line
remains at weld joint.

Formation of residual stress

(a) Shrinkage of unrestrained weld
(b) Shrinkage of restrained weld
Typical residual stress distribution
14 OPen INteractive Structural Lab
Factors affecting on Faigue of Welded Joints
Residual Stress
15

 Behavior of longitudinal residual stresses subjected to a nominal tensile

stress max in the welding direction
 Local plastic straining occurs and the stress in the weld remains at 
YS .When load is removed, the stress at the weld becomes  YS -max .

- YS
max
 YS -max max

 YS

 Mechanism of shake down :

• Residual stress at weld toe is usually in the level of yield stress.
• Residual stress at weld toe changes during cyclic loading.

15 OPen INteractive Structural Lab

Factors affecting on Faigue of Welded Joints
Residual Stress
16

 Tensile residual stresses can lead to a reduction of fatigue life.

 Tensile cyclic load 0~ max → shifted to  YS - max ~  YS (Tensile)
 Compressive cyclic load -  YS ~0→shifted to  YS -max ~  YS(Tensile).

Effective stress from superposition of

applied and residual stress
Superposition of applied stress
and residual stress

Fatigue test results for fillet joints

containing
16 high tensile residual stress
OPen INteractive Structural Lab
Factors affecting on Faigue of Welded Joints
In Practical Engineering in Ship and offshore field
17

 Residual stress effect in marine & offshore field?

• Varying stress amplitude
• High stress beyond yield stress due to stress concentration
• Shake down effect
• Quite complicated especially at bracket toe end
→ Residual stress disappears soon after operation
→ Usually not considered

 Mean stress effect in marine & offshore field?

• Beneficial effect due to compressive mean stress is included.
• Some offshore regulations neglects it in a conservative way.
• Mean stress is calculated in static loading condition.

17 OPen INteractive Structural Lab

Factors affecting on Faigue of Welded Joints
Size effects
18

 The effect of weld bead width is

analogous to attachment length.
 The effect of plate width is also significant.
 An appropriate correction factors to
experiment data relevant to other
dimensions.

Dimensions relevant to size effects in Influence of plate width on stress

fillet and butt weld joints 18 concentration
OPen INteractive factor
Structural Lab
Factors affecting on Faigue of Welded Joints
Thickness effects
19

 Fatigue strength decreases with plate thickness.

 There exists larger probability of weld flaws in thicker plate.
 The number and severity of flaws is likely increase with size.
 More restriction on thick plate during the fabrication process
 Smaller stress gradient over plate thickness in thicker plate 
larger high stress zone.

Thickness effect due to the stress gradient over

plate thickness
19 OPen INteractive Structural Lab
Ch.3 Factors affecting on Faigue of Welded Joints
Size effects
20

 S-N curve representing size effect

Influence of plate thickness Influence of attachment length

20 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Transverse butt welds
21

 Ductile failure occurs in the parent material. No reduction in strength caused

by the weld
 Under fatigue loading, stress concentration associated weld reduces fatigue
strength, the effect of stress concentration and weld toe intrusions.
 Complete removal of the excess weld metal and weld toe instructions by
machining or grinding.
 Variations in shape of weld profiles → Large variation of fatigue life.
 Welds which have a minimum of excess metal and a smooth transition at
the weld gives the highest fatigue strengths.

Single sided Double sided

Butt welding Butt welding

Fatigue failure from the toe of a transverse

butt weld
21 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Transverse butt welds (Both sides)- Misalignment
22

 Axial misalignment
 Local bending stress, a = nominal applied axial stress
3e
b   a  a   b   a (1 
3e 3e
), K g  (1  )
T T T

Nominal stress range

Kg=Stress concentration factor
 Angular misalignment 1° , Kg= 0.3

Local stress range Kg a

Effect of weld profile on fatigue

strength of transverse butt welds Fatigue test results from axially
22 misalignedStructural
OPen INteractive butt welds
Lab
FATIGUE OF WELDED JOINTS
Transverse butt welds – Single Sided butt welds
23

 Single sided welds : pipes or rectangular or circular

hollow section.
 Lack of penetration or unfavorable bead shape with a
bad profile → low fatigue strength
 Backing strip is effective.

Use of ceramic backing

Root cracks of butt welds
strip
23 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Transverse butt welds – Single Sided butt welds
24

 Permanent backing :
 integral with one of the members.
 crack initiation at the junction of the weld metal and backing strip.
 useful to tubular component.

To be machined out

Fatigue failure in transverse butt weld

made on backing strip
24 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Transverse butt welds – S-N Curve
25

Detail Sketch of detail 105 cycles 2x106

Transverse butt welds

355 MPa 155 MPa
with good profile

Transverse electron
beam butt weld 310 MPa 155 MPa

Transverse butt weld

with poorer profile
260 MPa 116 MPa
(including submerged
arc)

Transverse butt weld

on a permanent 260 MPa 115 MPa
backing strip

25 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Longitudinal butt Welds
26

 The excess weld metal lies parallel to the direction of applied load → no
stress concentration.
 Stop/start positions, where the electrode is changed in manual welds, the
ripples on the surface.
 Far less severe than that at the edge of the excess weld metal in transverse
direction.
 Side plates produce severe stress concentration.

Fatigue cracking at the attachment

butt welded to edge of stressed plate Fatigue cracking from weld ripples, (b)
26 Stop/start
OPen positions
INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Longitudinal butt Welds – S-N Curve
27

Detail Sketch of detail 105 cycles 2x106

Plain as-rolled steel

300 MPa 200 MPa
plate

Continuous automatic
350 MPa 160 MPa
longitudinal weld

Continuous manual
325 MPa 140 MPa
longitudinal weld

Longitudinal butt weld

on tack welded 280 MPa 125 MPa
backing strip

Intermittent longi.
Fillet weld 260 MPa 120 MPa

27 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Longitudinal butt Welds – S-N Curve
28

Detail Sketch of detail 105 cycles 2x106

Web/flange weld at
230 MPa 95 MPa
cope hole

Fillet attachment to
edges of stresses 180 MPa 70 MPa
members

Butt attachment to
edges of stresses 180 MPa 70 MPa
members

28 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connections
29

 Attachments give rise to a general stress concentration in addition to

local effect of the weld.
 Non-load-carrying fillet weld : attachment welds not designed to
transmit the loads.
 Load-carrying fillet weld : transmits load from one member to
another.
 Some loads will be transmitted through non-load carrying joint.

Fillet welded Lug : load carrying

attachment member to lifting load,
but non-load carrying
member to bending
stress

Non-load carrying Load carrying

Fillet weld Fillet weld
29 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
30

 Stress concentration increases with increase in attachment

length, ’L’, the thickness of longitudinal attachment or the width of
doubling plate.
 Increase in main plate thickness can lead to a reduction in fatigue
strength.
 Crack initiates at weld toe end and propagate through the plate.

Examples of non-load carrying fillet welded attachment

30 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
31

 Hard to detect through visual inspection, dye penetrant is

used.

Fatigue cracking at toe of transverse fillet weld (5.5 mm crack depth)

31 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
32

 Attachment with single fillet weld : fatigue crack initiation at crack

root.
 Weld root crack is unlikely to be detected before final failure. No
measures to improve the fatigue strength by treating the weld toe.

 To return the ends of the weld around the

edge of the attachment → little
improvement of fatigue strength but,
sealing against corrosion.

Transverse attachment Fatigue failure from end of

welded from one side 32 longitudinal fillet weld Structural Lab
OPen INteractive
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
33

 Continuation of the weld around the ends of the edge type of

attachment and across the edge of the main plate.
→ adversely reduce fatigue strength due to larger geometric
discontinuity.

Fatigue strength = 70 MPa at

2X106 cycle Fatigue strength = 52 MPa at
2X106 cycle

Fatigue cracking from end of fillet welded

attachment to edge of stressed plate

33 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
34

Detail Sketch of detail 105 cycles 2x106

Fillet or butt welded

stiffeners or
attachments to the 250 95
surfaces of stresses
member

Cover plates on beam

160 65
flanges

Web/flange weld at
230 95
cope hole

Butt welded
attachments to the
180 70
edges of stressed
member
34 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Non load carrying
35

 Intermittent fillet welds joining web to a flange in a beam : the

geometric stress concentration, fatigue strength is slightly high. =
120 MPa at 2X106 cycle.
 Web/flange fillet weld at cope hole : stress concentration is
increased.

Web

Fillet

Fatigue cracking from the end of a longi. Fatigue cracking from the end of a
intermittent web/flange fillet weld. web/flange fillet weld at cope hole.
35 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
36
Fillet Welded Connection – Load carrying fillet welds
 Design of load carrying fillet welded joints for static loads
• Weld length H is determined such that P/2AW< allowable design stress
• Fig (a) : 2AW/AP =1, the same stress in both the plate and the weld < allowable
tensile stress (H =0.7B)
• Fig (b) : P/(total weld length X 4 throat thickness ) < allowable shear stress
 However, behavior under Fatigue loading condition is more complicated.
Fatigue strengths are dependent on weld configurations and joint forms

σw=P/2AW

Aw

σP=P/AP
Ap
Fig (a) Transverse load-carrying fillet weld Fig (b) Longitudinal load-carrying fillet weld

OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
37

 The location of failure depends on the ratio leg length/plate thickness (H/B) .
 Fatigue strength at weld toe = 85 MPa at 2X106
at weld root = 57 MPa at 2X106
 For the same fatigue life at weld toe and weld root,

 p (toe ) 2 Aw 2 H 85 1.5 B
    1.5 H  1.06 B
 w( root ) AP B 57 2

σw=P/2Aw
 Optimum performance : enough
weld metal to ensure that Aw
failure would be form the weld
toe rather than weld root.

σP=P/AP
Ap

37 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
38

 Cruciform joint : two basic stress concentrations : at weld root and at weld
toes.
 Large fillet leg length (H1) : Larger notch stress at weld toes and smaller
stress on weld throat → crack initiation at weld toe.
 Small fillet leg length (H2) : Larger stress on the weld throat and smaller
stress at weld toe → crack initiation at weld root.

Fatigue cracking for different leg length

Two possible sites for fatigue cracking
38 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
39

 Partial penetration welds can achieve a given throat dimension with

a leg length smaller than would be required in normal fillet weld.
→ efficient way of improving fatigue performance.
 Full penetration weld eliminates stress concentration at the weld root
and reduce stress concentration at the toes.
→ However, the improvement is limited, increase of 10 MPa at 2X106

Partial penetration fillet weld Full penetration butt welded cruciform joint

39 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
40

 Example of full penetration welds in ship structures

40 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
41

 Misalignment of cruciform joint introduces secondary

bending stress and increases the stress range at the
weld toe and weld root.
 Less effect for weld root crack since it is closer to
neutral axis of the section w.r.t. the induced bending
moment.
 In real structure, the restraint of cross plate inhibits
bending due to misalignment. (no restraint in test)

Joints failing from the weld toe in the plate 41 Joints failing from
OPen the weld Structural
INteractive root in weld
Labthroat
FATIGUE OF WELDED JOINTS
Transverse butt welds (Both sides)- Misalignment
42

 Stress concentration factors for butt welds (DNV –RP-C203,2010,Section

3.1.2)

1/2(𝑇2 − 𝑇1 )
6(𝛿𝑚 + 𝛿𝑡 − 𝛿0 0.1𝑇1
𝐾𝑔 = 1 +
𝑇11.5
𝑇1 1 + 1.5 𝑇2
𝑇2
𝑇1

42 OPen INteractive Structural Lab

FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying
43

Detail Sketch of detail 105 cycles 2x106

Full penetration T or
360 95
cruciform joints

Fillet welded T or
250 85
cruciform joints

Lap joint with

250 85
transverse fillets

Weld throat failure in

transverse weld
154 57
(based on stress on
weld throat)
43 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Root Crack
44

 Design Chart for fillet and partial penetration welds(DNV –RP-

C203,2010,Section 2.8)
 Design should be performed such that fatigue cracking from the root
is less likely than from the toe.

Design chart for partial penetration fillet weld (DNV-RP-C203)

44 OPen INteractive Structural Lab
FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying fillet welds
45

 Lap joint using longitudinal fillet weld : less sensitive to weld size since the
changes of section at the weld toe and weld root are parallel with the
direction of applied stress.
 Weld ends induce stress concentration in both the main and cover plates.
 Weld on the edge is more damaging than weld on surface.
 Under the same nominal stress (BXWP=2BCPWCP), failure will initiate from a
weld end and propagate into the cover plate.

Fatigue failure from weld ends in cover plate

Weld
ends  Designed
BWP<2BCPWCP
Lap joint made using longi. fillet joint
 Continued weld
around the end
Joint45with weld continued around end
OPen INteractive of cover
Structural Labplate
Ch. 4 FATIGUE OF WELDED JOINTS
Fillet Welded Connection – Load carrying
46

Detail Sketch of detail 105 cycles 2x106

Lap Joint with weld
continued around end 235 85
of cover plate
Lap joint with longi.
fillet welds – crack 230 80
initiates on surface
Lap joint with longi.
fillet welds – crack
200 66
initiates on edge of
cover plate

Load carrying weld on

plate edge – not 150 50
welded around end

Load carrying weld on

plate edge – welded 125 44
around end
46 OPen INteractive Structural Lab
Design Regulations
47
What is fatigue failure?
 Before 70’s, fatigue strength is not a critical issue to be treated in ship design stage.
 Remarkable increase in use of high tensile steel (‘70~’80) → high applied stress →
frequent occurrence of fatigue crack → high repair cost
 The fatigue strength of unwelded components (base material) increases with material
yield strength but this is not the case with welded joints ← crack propagation of
crack-like flaws is dominant and the rate of crack growth is little affected by material
strength
 During 80’~90s, lots of research in fatigue strength → fatigue strength assessment is
mandatory, cap on the use of high tensile steel is placed by ship owner.
 The most common and important failure mode together with buckling.
Demand of fatigue strength assessment based on wave load analysis is increasing,
especially for high risk vessel like LNG carrier. ( Required fatigue life of LNG carrier is
40 years)
 Offshore structure requires higher fatigue life than commercial ship due to no
inspection after re-docking (Commercial vessel : 25 years. Offshore structure : 40 ~
200 years depending on the accessibility)

OPen INteractive Structural Lab

Design Regulations
48
Fatigue strength assessment

 Assessment of fatigue life under one cyclic load is simple since it

can be predicted from S-N curve.

 However, fatigue damage is accumulated over the entire lifetime of a

vessel, it is hard to analyze the load and stress history.

 Due to a huge number of welded joints in a vessel, an engineering

sense to screen critical joints which are prone to fatigue crack is
important.

 Fatigue analysis using finite element method : construction of

detailed fine mesh model and hot spot stress evaluation under
random wave load are sophisticated work.

OPen INteractive Structural Lab

Design Regulations
49
Fatigue Critical Area

OPen INteractive Structural Lab

Design Regulations
50
Fatigue Critical Area

OPen INteractive Structural Lab

Design Regulations
51
Fatigue Critical Area

OPen INteractive Structural Lab

Design Regulations
52
Fatigue Critical Area

OPen INteractive Structural Lab

Design Regulations
Fatigue strength assessment : LNG Carrier
53

Hydrodynamic
Full Stochastic
Analysis Local Zoom
FE analysis
Model

OPen INteractive Structural Lab

Design Regulations
Fatigue strength assessment : LNG Carrier
54

Hopper

OPen INteractive Structural Lab

Design Regulations
Fatigue strength assessment : FPSO
55

OPen INteractive Structural Lab

Design Regulations
S-N Curve – DNV RP –C203 vs DNV 30.7 Ship
56

Rule

OPen INteractive Structural Lab

Design Regulations
Nominal Stress based Approach : DNV RP –C203
57

OPen INteractive Structural Lab

Design Regulations
Nominal Stress based Approach : DNV RP –C203
58

OPen INteractive Structural Lab

Design Regulations
Nominal Stress based Approach : DNV RP –C203
59

OPen INteractive Structural Lab

Design Regulations
Nominal Stress based Approach : DNV RP –C203
60

OPen INteractive Structural Lab

Design Regulations
Nominal Stress based Approach : DNV RP –C203
61

OPen INteractive Structural Lab

Design Regulations
62
Nominal stress approach – S-N Curve : DNV RP –C203

 S-N Curves in seawater with cathodic protection

OPen INteractive Structural Lab

Design Regulations
63
Nominal Stress Approach-Combined stress : DNV RP –C203
 Design Chart for fillet and partial penetration welds
 Depending on the angle between principal stress and welding line, different
S-N curves are applied.

Classification E or F Classification C2

S-N curve depending on stress direction normal to

the weld and crack propagation
OPen INteractive Structural Lab
Design Regulations
64
Hot Spot Stress Approach

 Use of finite element analysis and difficulty in nominal stress → hot

spot stress for fatigue life assessment.
 Linear interpolation from t/2 and 3t/2
 Or σhot spot = Kg∙σnominal, Kg : structural stress concentration factor.

Schematic stress distribution at a hot spot

OPen INteractive Structural Lab
Design Regulations
Hot Spot Stress Approach – Modeling
65

 Shell element modeling :

• Modeling is easy but higher stress than the actual.
• 8-noded element is more flexible than 4-noded → less stress
• t x t mesh size
 Solid element modeling :
• Modeling is quite difficult but close to the actual.
• Fillet weld is modeled.
• 20-noded element is more flexible than 8-noded, t x t mesh size

Different hot spots Extrapolation in FE Extrapolation in FE

Shell Element Model Solid Element Model
OPen INteractive Structural Lab
Design Regulations
Hot Spot Stress Approach–Hot Spot Stress
66

 Derivation of stress at read out points 0.5 t and 1.5 t

 If mesh size is t x t in shell element, top surface stress to be read at
mid side nodes along A-B
 If solid element, stress to be extrapolated to the surface.
 If element size > t, fit a second order polynomial to the element
stresses in the three first elements and derive 0.5t & 1.5t stresses.

Extrapolated hot
spot stress

Intersection
line

Hot Spot

Gaussian integration point

Derivation of hot spot stress for
Example of derivation of hot spot stress element
OPensize larger than
INteractive t x tLab
Structural
Design Regulations
67
Hot Spot Stress Approach – S-N Curve : DNV RP-C203

 D Curve in air
Stress range (MPa)

Number of cycles
S-N curves in air
OPen INteractive Structural Lab
Design Regulations
Hot Spot Stress Approach – S-N Curve : DNV RP-C203
68

 Relationship between S-N Curves

D – curve, SCF=1.0 F1 curve, SCF=1.43

log N  12.164  3.0 log  log N  12.164  3.0 log( SCF ( 1.43)   )
 11.699  3.0 log 

OPen INteractive Structural Lab

Design Regulations
Hot Spot Stress Approach – S-N Curve : DNV CN.30.7
69

 Relationship between S-N Curves

OPen INteractive Structural Lab

Design Regulations
Hot Spot Stress Approach – S-N Curve : DNV CN.30.7
70

 Stiffener support is used to mitigate stress concentration at the intersection

between stiffener flange and web section .

OPen INteractive Structural Lab

Design Regulations
Hot Spot Stress Approach – S-N Curve : DNV CN.30.7
71

 Fatigue crack at lower hopper knuckles is caused by external dynamic

pressure and internal dynamic pressure of ballast tank or cargo tank.
 Fatigue crack initiates on inner bottom plate and propagates along the
welding line.

OPen INteractive Structural Lab

Design Regulations
Notch Stress Approach
72

 Effective notch stress is the total stress at the root of a notch, obtained assuming
linear-elastic material behaviour.
 For structural steels an effective notch root radius of r = 1.0 mm has been
verified to give consistent results.
 The method is restricted to welded joints which are expected to fail from the
weld toe or weld root and it is limited to thicknesses t 5 mm.
 Flank angles of 30° for butt welds and 45° for fillet welds are suggested.
 After certain post weld improvement procedures such as grinding, the actual
geometrical radius may be used in the effective notch stress analysis.

Frank angle and notch radius Analysis of effective notch stress

OPen INteractive Structural Lab
Design Regulations
Notch Stress Approach
73

 Calculation of effective notch stress by the finite element method : a fine

element mesh is used around the notch region.
 This maximum surface stress directly from the nodal stress calculated at
the surface or from extrapolation of element stresses to the surface.
 The effective notch stress to be used together with the recommended S-N
curve is the maximum calculated surface stress in the notch.

Notch stress based S-N 선도

OPen INteractive Structural Lab
Design Regulations
74
Fatigue Limit

 Fatigue limit is the maximum stress range below which the fatigue
strength is infinite.
 Permitted in design only when all stress range blocks are below the
fatigue limit.
 Otherwise, the slop is reduced from (-1/3) to (-1/5).

Stress cycling where further fatigue Stress cycling where a detailed

assessment can be omitted fatigue assessment is required

OPen INteractive Structural Lab

Design Regulations
75
Mean Stress Effect : DNV RP –C203
 Mean stress influence for non welded structure
 For fatigue analysis of regions in the base material not significantly
affected by residual stresses due to welding, the stress range may be
reduced if part of the stress cycle is in compression.
 t  0.6  c
fm 
t  c
  
 t  tension stress  max  static  ,0 
 2 
  
 c  compressive stress  min 0,  static  
 2  Stress Range Reduction Factor (fm)
log N  log a  m log( f m  )
 The calculated stress range may be multiplied with the reduction factor
fm before entering the S-N curve

 For welded joint

 Mean stress effect is neglected for fatigue assessment of welded
connections due to tensile residual stress around welded joint.
OPen INteractive Structural Lab
Design Regulations
76
Mean Stress Effect : DNV RP –C203
 min
Load Ratio R  Zero to max
 max

fm =1.0

Fully reversed
R=-1

Mean stress effect on S-N Curve

fm =0.8
Fatigue life is increased by
(1/0.8)3= 1.95 times

Zero to min

 max   min 1
m  fatigue life 
2 (f m  ) 3 fm =0.6
Fatigue life is increased by
Stress Range Reduction Factor (fm) (1/0.6)3= 4.63 times
OPen INteractive Structural Lab
Design Regulations
77
Mean Stress Effect : DNV 30.7 Ship Rule
 For Base material
  
 t  0.6  c  t  tension stress  max  static  ,0 
fm   2 
t  c   
 c  compressive stress  min 0,  static  
 2 
Stress Range Reduction Factor (fm)

 The calculated stress range may be multiplied with the reduction factor
fm before entering the S-N curve

 For welded joint

 A hot spot region is subjected to tensional residual stress.
 However, residual stresses due to welding and construction are reduced
over time as the ship is subjected to external loading.
 The mean stress effect is applicable. (Slightly larger than Base material)

 t  0.7  c
fm 
t  c
OPen INteractive Structural Lab
2 S-N Curve – DNV RP-C203 Offshore Rule
78
Thickness Effect
 Basic S-N curve
Tested plate thickness =22mm (DNV 30.7 Ship Rule)
Tested plate thickness =25mm (DNV RP-C203 Offshore Rule)

 Surface stress on the weld bead is increasing exponentially

 If plate thickness is beyond the reference thickness , modification of S-N curve is
needed.
k
 t 
log N  log a  m log   
t
 ref 
 t : plate thickness
 k : thickness exponent
DNV RP-C203 : k =0~0.3
DNV CN 30.7 : k=0.25

Stress distribution through thickness

OPen INteractive Structural Lab

Design Regulations
79
Effect of corrosive environment : DNV RP –C203
 Decrease in Fatigue Strength in Corrosive
environment.
 Cathodic protection : to control the corrosion of a
metal surface by connecting the metal to be
protected with another more easily corroded
"sacrificial metal" to act as the anode of the
electrochemical cell.

S-N curves in air S-N curves in seawater with cathodic protection

OPen INteractive Structural Lab

Design Regulations
80
Nominal Stress based Approach : DNV RP –C203

 S-N Curves in air

S-N Curves in seawater with cathodic protection

OPen INteractive Structural Lab

Design Regulations
81
Effect of corrosive environment : DNV CN 30.7
 Ship rule is not conservative.
 SN-Curve for air = SN-Curve for cathodic protection.

OPen INteractive Structural Lab

Calculation of Fatigue life
82
Palmgren-Miner rule
 EX) During 10 years, 1 : n1 cycles
2 : n2 cycles
3 : n3 cycles
k
 Damage ratio D 
N
ni
i 1 i

D = n1/N1 + n2/N2 + n3/N3

Fatigue Life= (n1+ n2 + n3)/D = 10 years/D
OPen INteractive Structural Lab
Calculation of Fatigue life
83
Palmgren-Miner rule
 Palmgren-Miner Rule
k k
ni n n n
D   Di    1  2    k
i 1 i 1 N i N1 N 2 Nk

Where,
k= number of stress blocks
ni = number of stress cycles in stress block with constant stress range

Ni = number of cycles to failure at constant stress range

L0
L = fatigue life L
D k
Lo = the time for the total number of stress cycles n 0   ni
i 1

OPen INteractive Structural Lab

Calculation of Fatigue life
84
Example 1
 Calculate fatigue life of the following butt weld using E curve subjected to
long term stress range in sea water in cathodic protection. Use DNV-RP-
C203. Plate thickness = 40 mm. Mean stress is tensile stress.

 i

 i

OPen INteractive Structural Lab

Calculation of Fatigue life
85
Example 1 : Solution
 It belongs to E class and thickness effect to be considered.
 E class
log a  11.61, m  3 for N  106
log a  15.35, m  5 for N  106
for N  106
t 0 .2
log N  log a  m log( )  m log 
25
40
11.61  3.0 log( ) 0.2  3.0 log 
25
 11.49  3.0 log 
for N  106
40 0.2
log N 15.35  5.0 log( )  5.0 log 
25
 15.15  5.0 log 

Damage ratio = 0.247

 Fatigue Life = n0/0.247 =1,664,413
OPen INteractive Structural Lab
Calculation of Fatigue life
86
Equivalent Stress Range
 The same tress range is assumed to be applied across the entire life
and it gives the same damage ratio subjected to variable stress
ranges

n  n0  eqm k
, n0   ni
k k m
n
D i  i , N i  a  i m
i D
a i 1
i 1 N i i 1 a

1/ m
 k ni  im 
 eq   
 i 1 n0 

OPen INteractive Structural Lab

Post weld improvement
87
Weld Toe grinding - mitigation of stress concentration

 A major cause of fatigue damage in welded structures : stress

concentration at the toe + crack-like flaws.
 Machining and grinding to eliminate such flaws and give a smoother
profile → improvement fatigue strength.
 Disc grinding is completely more quickly.
 Hand-held burr grinding is more effective.

Burr machining and disc grinding

OPen INteractive Structural Lab

Post weld improvement
88
Weld Toe grinding – mitigation of stress concentration
 Fatigue life improvement is valid only when corrosion protection is
effective
 After protection, corrosion induces notches.
 Adverse effect of strength reduction due to the decrease in area.

As Welded Burr Ground

OPen INteractive Structural Lab
Post weld improvement
89
Weld Toe grinding – side effect, corrosion

 Corrosion pitting of the ground metal surface virtually eliminates the

benefit of burr grinding.
 The ground surface must be adequately protected by cathodic
protection system or permanent protection means like a paint
system.

Corrosion pitting

OPen INteractive Structural Lab

Post weld improvement
90
Weld Toe grinding – mitigation of stress concentration
 Necessary to remove all traces of the original weld toe and material to a
depth of 0.5~1.0 mm below any undercut.
 0.5mm <Depth of grinding < 2mm or 7% of plate thickness

Depth of grinding Depth measurement gauge

OPen INteractive Structural Lab

Post weld improvement
TIG dressing
91

 TIG and plasma dressing

 As effective as local grinding but considerably faster.
 To re-melt the weld toe and wash the weld pool into the plate surface as to
produce a smoother weld profile and remove inherent flaws.
 TIG dressing calls for precise positioning of the arc
 Plasma technique gives a larger area of heating → less demanding

Weld cross section Influence of electrode position fatigue

91 strength
OPenofINteractive
plasma dressed filletLab
Structural welds
Post weld improvement
Weld Toe grinding - Burr Grinding
92

 A high speed pneumatic, hydraulic or electric grinder : rotational

speed :15,000 ~ 40,000 rpm.
 The diameter :10 ~ 25 mm for 10t ~50t.
root radius > 0.25t.
 The high-speed grinding tool removes material at a high rate.
 The cutting operation itself produces hot, sharp cuttings and some noise.
 Heavy protective clothing together with leather gloves, safety glasses and
ear protection are mandatory

Example of protective clothing

Pneumatic grinder and burrs used during weld toe burr grinding.
OPen INteractive Structural Lab
Post weld improvement
Hammer peening - compressive residual stress
93

 Effective stress = applied stress + residual stress

 Tensile residual stress + compressive applied stress = no compressive
mean stress effect
 Compressive residual stress + tensile applied stress = compressive mean
stress effect

 High compressive stresses at the site of stress

concentration leads to an improvement in
fatigue strength.
 Cold working the material surface →
compressive stress on surface which is
balanced by a residual tensile stress within
core.

Superimposition of applied
and residual stresses OPen INteractive Structural Lab
Post weld improvement
Hammer peening - compressive residual stress
94

 Hammer peening
 pneumatic or electric hammer with round-ended tool → compressive stress
on the surface of stress concentration
 Noise problem : sometimes not allowed for health reasons → Needle
peening

Hammer peeing fillet weld toe Needle peeing fillet weld toe Weld cross-section

OPen INteractive Structural Lab

Post weld improvement
Hammer peening - Equipment
95

 Suitable pneumatic hammer gun has a 15 to 30 mm diameter piston,

operates at an air pressure of 5 to 7 bars and delivers 25 to 100
blows per second. Impact energy is typically in the range 5 to 15
Joules.
 The weight of the gun is from about 1 to 3.5 kg.
 Hammer peening have made use of hammer gun, both of which are
primarily intended for use as chipping (쪼다) hammers.

Pneumatic riveting guns used for

hammer peening.
OPen INteractive Structural Lab
Post weld improvement
Hammer Peening - Procedure
96

OPen INteractive Structural Lab

Post weld improvement
Effects of Post Weld Improvement
97

 Comparison of post weld improvement methods

Improvement of S-N curve

OPen INteractive Structural Lab

Post weld improvement
98
Improvement of Fatigue Life by Fabrication DNV RP-C203
 Grinding should extend below the plate surface by a rotary burr.
 The grinding depth should not exceed min (2 mm, 7% of the plate
thickness).
 A good design practice to exclude this factor at the design stage and
to keep the possibility of fatigue life improvement as a reserve.
 Crack grows faster after initiation → shorter inspection intervals
during service life in order to detect the cracks before they become
dangerous.

Grinding of welds OPen INteractive Structural Lab

Post weld improvement
99
Improvement of Fatigue Life by Fabrication DNV RP-C203
 Improvement on fatigue life by different methods

OPen INteractive Structural Lab

Calculation of Fatigue life -Post weld improvement
Example 2
100

 The following welded joint is subjected to cyclic load which gives the
following nominal stress range during 10 years in seawater with
cathodic protection in accordance with DNV RP-C203

Applied nominal Number of

stress range (MPa) applied cycles
70 5.0X105
80 4.0X105
90 3.0X105
100 2.0X105
110 1.5X105

 Principal stress direction (f) = 40 degrees

 Plate thickness = 30 mm
 Static nominal mean stress : 20 MPa (Tensile stress)
 Steel : High Tensile steel (Yield stress = 355 MPa)
 Weld toe grinding is applied.

OPen INteractive Structural Lab

Calculation of Fatigue life -Post weld improvement
Example 2 : Solution
101

 Principal stress direction (f) = 40 degrees  Plate thickness = 30 mm

t>25
30
45   t  0.20 
60
log N  log a  m log    
  25  
 
  t  0.20 
log N  11.610  3.0 log     N  106
  25  
 
  t  0.20 
log N  15.350  5.0 log     N  106
  25  
 

OPen INteractive Structural Lab

Calculation of Fatigue life -Post weld improvement
Example 2 : Solution
102

 E Curve
Stress range (MPa)

Number of cycles

OPen INteractive Structural Lab

Calculation of Fatigue life -Post weld improvement
Example 2 : Solution
103

 Static nominal mean stress : 20 MPa (Tensile stress)

→ Mean stress effect is not allowed in DNV RP-C203.

 Fatigue Damage Calculation

Δσnominal ni Ni Di Remarks
6 For N  106
70 500,000 1,110,014 0.45044 Use N>10 curve
6   t  0.20 
80 400,000 713,216 0.56084 Use N<10 curve log N  11.610  3.0 log    
  25  
90 300,000 500,915 0.59890 Use N<106 curve  
100 200,000 365,167 0.54769 Use N<106 curve For N  10 6

110 150,000 274,355 0.54674 Use N<106 curve   t  0.20 

log N  15.350  5.0 log    
  25  
D= 2.705  

Fatigue Life = 10 year / 2.705 = 3.7 year

 Steel : High Tensile steel (Yield stress = 355 MPa)

 Weld toe grinding is applied.

Post Improvement : yield stress > 350 MPa , Improvement on fatigue life = 3.5
Fatigue Life = 3.7 year X 3.5 = 13.0 years
OPen INteractive Structural Lab
Calculation of Fatigue life -Post weld improvement
Example 3
104

 The following welded joint is subjected to cyclic load which gives the
following hotspot stress range during 20 years in seawater with
cathodic protection in accordance with DNV CN-30.7

 Plate thickness = 30 mm
 Static nominal mean stress : 20 MPa (Tensile stress)
 Steel : High Tensile steel (Yield stress = 355 MPa)
 Weld toe grinding is applied.

OPen INteractive Structural Lab

Calculation of Fatigue life -Post weld improvement
Example 3 : Solution
105

 Static nominal mean stress : 20 MPa (Tensile stress)

 t  0.7  c  t  tensile stress  max( static   / 2,0)
fm 
t  c  c  compressive stress  min(0,  static   / 2)

 Fatigue Damage Calculation

For N  106
  t  
0.25

log N  12.164  3.0 log  f m    

  25  

For N  106
  t  
0.25

log N  15.606  5.0 log  f m    

  25  


Fatigue Life = 20 year / 2.63 = 7.61 year

 Steel : High Tensile steel (Yield stress = 355 MPa)

 Weld toe grinding is applied.

Post Improvement : yield stress > 350 MPa , Improvement on fatigue life = 3.5
Fatigue Life = 7.61 year X 3.5 = 26.63 years

OPen INteractive Structural Lab