dnvgl-cg-0129 fatigue assessment of ship structures...changes - current class guideline —...

The electronic pdf version of this document available free of chargefrom httpwwwdnvglcom is the officially binding version

DNV GL AS

CLASS GUIDELINE

DNVGL-CG-0129 Edition October 2015

Fatigue assessment of ship structures

FOREWORD

DNV GL class guidelines contain methods technical requirements principles and acceptancecriteria related to classed objects as referred to from the rules

copy DNV GL AS

Any comments may be sent by e-mail to rulesdnvglcom

If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of DNV GL then DNV GL shallpay compensation to such person for his proved direct loss or damage However the compensation shall not exceed an amount equal to tentimes the fee charged for the service in question provided that the maximum compensation shall never exceed USD 2 million

In this provision DNV GL shall mean DNV GL AS its direct and indirect owners as well as all its affiliates subsidiaries directors officersemployees agents and any other acting on behalf of DNV GL

Cha

nges

- c

urre

nt

Class guideline mdash DNVGL-CG-0129 Edition October 2015 Page 3Fatigue assessment of ship structures

DNV GL AS

CHANGES ndash CURRENT

This is a new document

Con

tent

s


DNV GL AS

CONTENTS

Changes ndash current 3

Section 1 General 121 Introduction122 Symbols and abbreviations12

21 Symbols 1222 Abbreviations 1923 Coordinate system and sign conventions 20

3 Application2131 General 2132 Calculated and observed fatigue strength 2233 Vibrations 2234 Low cycle fatigue 22

4 Methods for fatigue assessment 2241 Stress types used for fatigue assessment 2242 Methods for stress calculation2243 Selection of prescriptive or direct methods for calculation of stresses23

5 Fatigue approaches2451 General 2452 Crack failure modes2453 Assessment using local and nominal stress approach 2654 Assessment using hot spot stress approach 2755 Stress range 28

6 Additional class notations2861 General 2862 CSA 2863 RSD2864 PLUS2965 HMON 2966 VIBR and COMF-V 2967 NAUT and routing2968 WIV 29

7 Definitions 29

Section 2 Fatigue Capacity 331 Introduction33

Con

tent

s


DNV GL AS

2 S-N curves 3321 General 3322 S-N curves and detail classification3323 S-N curves for in-air environment 3424 S-N curves for stress along the welds3725 Nominal stress S-N curves 3926 Other steel types 39

3 Corrosive environment3931 General 3932 Fatigue assessment in corrosive environment39

4 Mean stress effect 3941 Base material and free plate edges 4042 Welded joints40

5 Thickness effect4151 General 4152 Effective thickness of welds 4453 Thickness effect of welds in prescriptive analysis of longitudinal end

connections 4554 Thickness effect of welds in FE analysis 4555 Thickness effect of base material 46

6 Material factor 46

Section 3 Fatigue strength representation 471 Introduction472 Fatigue stress range47

21 General 4722 Scantlings approach factor fc4723 Environmental factor fe4824 Post-weld treatment 4825 Fatigue stress range 48

3 Fatigue damage and fatigue life calculation 4931 Fatigue damage accumulation with Palmgren - Minerrsquos rule 4932 Time in air and corrosive environment 5033 Annual fatigue damage 5034 The combined fatigue damage 5135 The total fatigue damage from multiple loading conditions 5136 Fatigue life 52

4 Permissible stress range 535 Permissible stress concentration factor and required FAT class 54

Con

tent

s


DNV GL AS

Section 4 Prescriptive fatigue strength assessment 551 Introduction55

11 Stress approach for longitudinal end connections5512 Calculation procedure for longitudinal end connections 5513 Definition of stress components 5614 Calculation of stress components 57

2 Wave induced loads583 Fatigue stress range58

31 Hot spot stress or nominal stress5832 Prescriptive fatigue stress range5833 Stress range 5934 Mean stress59

4 Global hull girder stress 6041 Stress due to wave induced hull girder loads 6042 Hull girder vibrations 6143 Stress due to still water hull girder bending moment61

5 Stress due to bending of PSM6251 General 62

6 Local stiffener bending stress6261 Wave induced stiffener bending stress 6262 Still water stiffener bending stress 64

7 Local relative deflection stress 6571 General 6572 Relative deflection definition 6573 Sign convention 6574 Estimate of relative deflection6575 Prescriptive stress due to relative deflection derived by FE analysis 6576 Stress due to relative displacement in still water 6777 Consideration of reduced local stiffener bending stress 67

8 Stress concentration factors6881 Longitudinal stiffener end connections6882 Other connection types and overlapped connections68

Section 5 Direct Fatigue strength assessment691 General 69

11 Definition6912 Stress transfer functions 6913 Assumptions 69

Con

tent

s


DNV GL AS

14 Hydrodynamic theory692 Fatigue strength assessment70

21 Fatigue damage calculations 703 Component stochastic analysis 70

31 Load components 7132 From load to stress transfer functions7233 Stress factors per unit load 7234 Splash zone correction7335 Double hull and relative deflection stress 74

4 Full stochastic analysis7441 Introduction7442 Assessment of local details76

Section 6 Finite Element Analysis 771 Introduction77

11 Thickness 7712 Application7713 Stress to be used from finite element analysis 7814 Stress field at a welded detail 7915 Fatigue stress range based on EDW approach 7916 Fatigue stress range based on direct calculations8117 Hot spot S-N curve 82

2 Finite element modelling 8221 General 8222 FE models 8223 Load application 8224 Mesh size 8325 4-node or 8-node elements 8326 Base material free plate edges8327 Example Hatch corners and hatch coaming end bracket 83

3 Derivation of hot spot stress 8531 General 8532 Read out point at t2 8533 Derivation of stress at read out point t2 8534 Reduced hot spot stress for highly localized stress peaks with plate bending 87

4 Procedure for analysis of standard details8841 Procedure for analysis of hot spot stress at weld toe of welded details 8842 Procedure for analysis of local stress of base material at free plate edge88

Con

tent

s


DNV GL AS

5 Procedure for analysis of hot spot stress of web-stiffened cruciformjoints88

51 General 8852 Calculation of hot spot stress at the flange8953 Calculation of hot spot stress in way of the web92

6 Procedure for analysis of hot spot stress of simple cruciform joints 9361 Limitation to hot spot stress approach 9362 Correction of hot spot stress approach for simple cruciform joints 94

Section 7 Improvement of Fatigue Life by Fabrication 951 Introduction95

11 General 952 Grinding 95

21 Weld toe burr grinding9522 Flush grinding of butt welds 9723 Fatigue life improvement factor for burr grinding97

Section 8 Detail design standard991 Introduction992 Detail design standard for longitudinal end connections without top

stiffener9921 Design standard A - slots with and without lug plate9922 Design standard B - collar plate 10123 Equivalent design of longitudinal end connections without top stiffener102

3 Detail design standard for different ship types104

Section 9 References1051 References 105

Appendix A Stress Concentration Factors and FAT classes 1071 General 107

11 Definition of stress concentration factors 10712 Determination of stress concentration factors 10713 Definition of FAT class10714 Basis108

2 Longitudinal end connections 10821 Tabulated K factors10822 Soft toe of web stiffener and backing bracket11823 Unsymmetrical stiffener flange 120

3 Butt welds 1234 Flange connections 127

Con

tent

s


DNV GL AS

5 Welded attachments 13251 Attachments welded to a plate or stiffener13252 Termination of stiffeners13453 Doubling plates 135

6 Simple cruciform joints1367 Scallops and block joints 1388 Lower hopper knuckles1399 Pipe connections14110 Holes with edge reinforcements 14111 Rounded rectangular elliptical and oval holes162

Appendix B Fatigue capacity 1671 Failure criterion inherent in the S-N curves1672 Other S-N curves 167

21 S-N curves for base material of high strength steel above 500Nmm2 16722 Qualification of new S-N curves based on fatigue test data168

3 Effect of fabrication tolerances1694 Design chart for fillet and partial penetration welds169

Appendix C Fatigue damage calculations and fatigue stress design tables 1711 Weibull long term stress distribution1712 Fatigue damage accumulation 171

21 Practical damage calculation with one slope S-N curve17122 Closed form damage estimate for one-slope S-N curves and Weibull

distribution 17223 Closed form damage estimate for two-slope S-N curves and Weibull

distribution 17324 Closed form damage estimate for one slope S-N curve and short term

Rayleigh distribution 17325 Closed form damage estimate for two-slope S-N curve and short term

Rayleigh distribution 1743 Maximum allowable stress range1744 Guidance for omission of fatigue analysis177

Appendix D Prescriptive fatigue strength assessment1791 Bending stress of PSM179

11 Prescriptive double hull bending stress at transverse bulkhead17912 Prescriptive wave induced double side bending stress at the transverse

bulkhead 18013 Prescriptive still water double side bending stress at the transverse

bulkhead 1812 Local relative displacement stress181

Con

tent

s

Class guideline mdash DNVGL-CG-0129 Edition October 2015 0Fatigue assessment of ship structures

DNV GL AS

21 Prescriptive wave induced relative deflection for double hull vessels18222 Prescriptive relative deflection for double hull vessels due to static pressure 184

3 Local plate bending stress18531 Wave induced plate bending stress 18532 Still water plate bending stress 186

Appendix E Finite element analysis1871 FE modelling 187

11 Mesh size 18712 Solid elements 187

2 Derivation of hot spot stress 18721 Extrapolation from t2 and 3t2 18722 Derivation of effective hot spot stress considering principal stress

directions1883 Derivation of stress at read out points 189

31 Derivation of stress at read out points t2 and 3t21894 Procedure for analysis of hot spot stress for tuned details 193

41 Bent hopper tank knuckles 1935 Screening procedure195

51 Introduction 19552 Procedure 19553 Screening fatigue criteria19654 Stress read out procedure for bracket toe19655 Stress read out procedure for knuckle detail 19756 Read out point stress198

6 Derivation of stress concentration factors for alternative endconnection designs199

7 Verification of analysis methodology 200

Appendix F Workmanship and link to analysis procedures 2011 Tolerance limits 2012 Weld profiling by machining and grinding 2023 TIG dressing 2034 Hammer peening 2035 Improvement factors for TIG dressing and Hammer peening203

Appendix G Uncertainties in Fatigue Life Predictions 2051 Introduction2052 Probability of fatigue failure2063 Size of defects inherent in S-N data and requirements to non-

destructive examination209

Con

tent

s


DNV GL AS

Appendix H Low cycle fatigue 2111 General 2112 Number of static design cycles and fraction in different loading

conditions2123 Loading conditions213

31 Combination of static stress21332 Pressure loads219

4 Simplified calculations of stresses21941 Hot spot stress range due to wave loading 21942 Hot spot stress range for low cycle fatigue 22043 Combined hot spot stress range 22044 Plasticity correction factor22145 Total elastic stress component223

5 Fatigue damage calculations for low cycle fatigue2246 Correction of the damage or stress 2247 Combined fatigue damage due to high and low cycle fatigue 2258 Example of application 226

Appendix I Wave induced vibrations for blunt vessels 2291 Introduction2292 How to include the effect of vibration 2303 Vibration factor for blunt vessels 2304 Application of the vibration factor fvib 2315 Effect of the trade 2326 Model tests procedure 2337 Numerical calculation procedure2338 Full scale measurements 233

Appendix J Considerations of special details2351 Crane foundations and foundations for heavy top side loads 235

11 Introduction 23512 Fatigue assessment based on wave induced loads235

Sec

tion

1


DNV GL AS

SECTION 1 GENERAL

1 IntroductionThis Class Guideline (CG) gives a transparent and detailed description of the fatigue assessment methodsand procedures to support and satisfy the requirements given in the rules The rules provides therequirements and is governing compared to this CG

2 Symbols and abbreviations

21 SymbolsThe symbols in Table 1 are used in this CG For symbols not defined in this CG reference is made to RU SHIPPt3 Ch1 Sec4 and RU SHIP Pt3 Ch9

Table 1 Symbol list

Symbol Meaning Unit

α Angular misalignment degrees

α(j) Fraction of time in loading condition j compared to other loading conditions -

αh Angle between main plate and attachment plate of hopper knuckles degrees

β Relative wave heading to bow direction degrees

βs Fraction of permissible still water bending moment -

βh Hot spot stress correction factor for hopper knuckles -

γ Incomplete gamma function -

Γ( )

Γ( )

Complete gamma functionComplementary incomplete gamma function

-

δ Standard deviation of S-N curve -

δFwdik(j)δAftik(j) Relative deflection between transverse bulkhead and adjacent flange aft and forward ofthe transverse bulkhead

mm

Δσ Stress range S-N curve variable nominal or local stress range Nmm2

Δσ0 Reference (largest) stress range value exceeded once out of n0 cycles Nmm2

ΔσBS Local stress range at free plate edge from FE analysis or beam theory Nmm2

ΔσFS(j) Fatigue stress range for loading condition j ie the stress range to enter into the S-N curve Nmm2

ΔσFS1ΔσFS2 Fatigue stress range for stress normal and parallel to the weld Nmm2

ΔσHS Hot spot stress range from FE analysis or beam theory Nmm2

ΔσΗS1 ΔσΗS2 Hot spot stress range for stress normal and parallel to the weld Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

ΔσG Global stress range Nmm2

ΔσL Local stress range Nmm2

Δσloc Local stress range at based material or free plate edge Nmm2

Δσh Nominal stress range due to horizontal bending Nmm2

Δσn Nominal stress range (giving a total damage D = 1) at 10-2 probability level Nmm2

Δσperm Permissible stress range giving a total damage D = 1 Nmm2

Δσq Reference stress range at knuckle of the S-N curve ie at N=107 cycles Nmm2

Δσv Nominal stress range due to vertical bending Nmm2

ξ Weibull shape parameter (ξi) -

η Stress magnification factor -

η Fatigue usage factor -

ηWηldηbd

ηSηlsηbs

Pressure normal coefficients for dynamic sea liquid and dry bulk cargo pressure

Pressure normal coefficients for static sea liquid and dry bulk cargo pressure

-

θ Toe angle (relevant for soft toe definition) degrees

q Angle between local finite element x-direction and local X-direction of principal stress rad

q Roll angle rad

μ Parameter in damage calculations accounting for change of inverse slope of S-N curve -

ν Parameter in the incomplete gamma distribution -

ν0 Average zero-crossing frequency of the stress (based on the wave bending moment) Hz

ρ Density of liquid tonsm3

σ Stress amplitude surface principal stress at hot spot Nmm2

σbending Stress amplitude from the bending component from FE analysis at the element surface Nmm2

σcomb Combined transverse stress and shear stress of fillet welds on holes with reinforcements Nmm2

σd Local relative deflection stress between framesbulkheads Nmm2

σdD Dynamic wave induced stress amplitude from relative deflection Nmm2

σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)

Relative deflection stress at location lsquoarsquo or lsquofrsquo aft or forward of the transverse bulkhead forload case i snap shot k and loading condition j

Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

σdh Stress amplitude due to bending of PSM eg bending of double hull or framegirdergrillage system

Nmm2

σdhD Dynamic wave induced stress amplitude from double hull bending Nmm2

σG Global hull girder stress amplitude Nmm2

σGD Dynamic wave induced stress amplitude from global loads Nmm2

σL Local stiffener bending stress amplitude Nmm2

σLD Dynamic wave induced stress amplitude from local loads Nmm2

σdS Static stress amplitude from relative deflection Nmm2

σdhS Static stress amplitude from double hull bending Nmm2

σGS Static stress amplitude from global loads Nmm2

σHS Hot spot stress to be used in fatigue assessment by hot spot stress approach Nmm2

σLS Static stress amplitude from local loads Nmm2

σloc Local stress from FE analysis eg base material free plate edge of hatch corner notexcluding welded details

Nmm2

σmdD Mean stress of dynamic wave induced stress from relative deflection Nmm2

σmdhD Mean stress of dynamic wave induced stress from bending of PSM Nmm2

σmLD Mean stress of dynamic wave induced stress from local loads Nmm2

σmGD Mean stress of dynamic wave induced stress from global loads Nmm2

σmean Mean stress amplitude Nmm2

σmean i(j)pX

σmean i(j)pY

Mean stress amplitude corrected to the principal stress direction of the hot spot lsquonormalrsquoand lsquoparallelrsquo to the weld

Nmm2

σmembrane Stress amplitude from the membrane component from FE analysis at the mid layer of theelement

Nmm2

σmax Maximum stress amplitude (positive is tension) Nmm2

σmin Minimum stress amplitude (negative is compression) Nmm2

σn Nominal stress in Nmm2 to be used in fatigue assessment by nominal stress approach Nmm2

σnotch Notch stress which is implicitly included in the hot spot S-N curves Nmm2

σna Nominal axial stress at from axial force at the stiffener flange Nmm2

σnb Nominal bending stress from unit lateral pressure at the stiffener flange in way of the hotspot

Nmm2

σHSa Hot spot stress determined at the stiffener flange from the axial load Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

σHSb Hot spot stress determined fat the stiffener flange from unit pressure Nmm2

σshift Surface principal stress amplitude at shifted position from the FE analysis Nmm2

σsurface Hot spot stress amplitude from the surface component from FE analysis Nmm2

σP Local plate bending stress amplitude Nmm2

σpX Principal stress amplitude in local X-direction Nmm2

σpY Principal stress amplitude in local Y-direction Nmm2

σ2 Secondary stress amplitude Nmm2

σ3 Tertiary stress amplitude produced by bending of plate elements between longitudinal andtransverse framesstiffeners

Nmm2

τ Shear stress amplitude Nmm2

Δτ Shear stress range Nmm2

Γ( )

Γ( )


-

ϕ Toe angle degrees

ω Wave frequency rads

ωi-n50 Net sectorial co-ordinate of the point being considered in the transverse cross section m2

Ak Stress to load ratio for load component k (unit depends on the load component) -

beff Effective flange of attached plating at the end of the span mm

bf-n50 Net flange width mm

bf Gross flange breath mm

bg-n50 Eccentricity of the stiffener equal to the distance from the flange edge to the centre line ofthe web

mm

bg Gross eccentricity mm

c Correction coefficient for bending stress when relative deflection is included ca or cf aft andforward of the bulkhead when the spans aft and forward differ

-

d Attachment length or top stiffener breath mm

d Weld width of butt welds attachments and cruciform joints (including attachmentthickness)

mm

d Burr grinding depth mm

D Total predicted fatigue damage -

D(j) Damage during the design fatigue life in loading condition j assuming 100 time at sea -

DE(j) Annual fatigue damage in air for loading condition j assuming 100 time at sea -

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

Dair Annual fatigue damage for all loading conditions in air including port time -

e Eccentric misalignmentdeflection mm

f0 Factor for fraction of time at sea -

fc Scantling approach factor for correction of stress according to net scantlings approach -

fd Reduction factor of wave fatigue moment for location x compared to maximum wavefatigue moment along hull girder

-

fe Environmental reduction factor -

fmaterial Correction factor for material strength -

fmean Mean stress reduction factor -

fN Correction factor for spectrum shape and number of load cycles -

fNL Correction factor for the effect on non-linearity -

fthick Factor for thickness effect gt=1 -

fvib Correction factor for wave induced vibration -

fw Post-weld treatment factor reducing the fatigue stress -

FAT Xreq Required FAT class giving a total damage D=1 Nmm2

hstf hstf-n50 Stiffener height and net web height including face plate mm

htoe Height of bracket toe (relevant for soft nose definition) mm

hHS Height from plate to hot spot on stiffener without top stiffener on the stiffener flange mm

hw Water head equivalent to the pressure at waterline m

hw-n50 Height of stiffener web without flange thickness mm

hHS Height from plate to hot spot on the stiffener web mm

IFwd-n50

IAft-n50

Area moment of inertia for longitudinal stiffeners aft or forward of the transverse bulkhead cm3

K Total stress concentration factor -

Ka Geometrical stress concentration factor for axial loading -

Kb Geometrical stress concentration factor for bending loading -

Kd Factor for relative deflection -

K1 S-N curve parameter (mean) ldquocrossing of x-axisrdquo upper part (Nmm2)m

K2 S-N curve parameter (mean minus two standard deviations) ldquocrossing of x-axisrdquo upperpart

(Nmm2)m

K3 S-N curve parameter (mean minus two standard deviations) ldquocrossing of x-axisrdquo lowerpart

(Nmm2)m

Kg Geometric stress concentration factor -

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

Kn Stress concentration factor due to shear bending of unsymmetrical stiffeners with lateralloading

-

KP Stress reduction factor for stress parallel with weld direction -

Kperm Permissible stress concentration factor based on the nominal stress approach giving D=1 -

Kmα Stress concentration factor due to angular misalignment -

Kme Stress concentration factor due to eccentric misalignment (normally plate connections) -

Km Stress concentration factor due to misalignment -

Kw Weld shape factor normally taken as 15 and included in the hot spot S-N curve -

ℓFwd ℓAft Stiffener span aft and forward of the transverse bulkhead m

ℓbdgℓbdg Fwd

ℓbdg Aft

Effective bending span of stiffener Aft and Forward of the transverse bulkhead m

log( ) 10th logarithm

ln( ) Natural logarithm

m S-N curve parameter for inverse slope -

Δm S-N curve parameter change of inverse slope for lower part of two slope S-N curvecompared to upper part

-

m0inj Zero spectral moment of stress response in short term condition (in) and loading conditionj

Nmm2

MwvLCik Vertical wave bending moment for the load case i and snap shot k (k=12) kNm

Mwh-LCik Horizontal wave bending moment for the load case i and snap shot k (k=12) kNm

Mσ-wt-LCik Dynamic bi-moment related to warping stress for the load case i and snap shot k (k=12) kNm2

Msw Permissible still water bending moment kNm

n Exponent in calculation of the thickness effect -

n Number of load cycles at a given probability of exceedance -

n Number of load components -

n0 Total number of cycles corresponding to the reference stress range level Δσ0 -

ni Number of load cycles in stress range block i -

nLC Number of applicable loading conditions -

ntot Total number of stress range blocks i -

ND Number of stress cycles per year -

NR Average number of cycles between each exceedance at a given probability of exceedanceeg NR=100 for 10-2 probability of exceedance

-

Ni Number of cycles to failure in stress range block i according to the S-N curve -

N Number of cycles to failure according to the S-N curve -

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

Pbdik(j) Dynamic dry bulk cargo pressure for load case i snap shot k and loading condition j kNm2

Pbs(j) Static dry bulk cargo pressure for loading condition j kNm2

Pldik(j) Dynamic liquid tank pressure for load case i snap shot k and loading condition j kNm2

Pls(j) Static liquid tank pressure for loading condition j kNm2

PS(j) Static sea pressure for loading condition j kNm2

PW Dynamic wave pressure kNm2

PWWL Dynamic wave pressure at water line kNm2

PWik(j) Dynamic wave pressure for load case i snap shot k and loading condition j kNm2

PWSZ Dynamic wave pressure in the splash zone kNm2

q Weibull scale parameter Nmm2

Q (Δσ) Probability level for exceedance of stress range Δσ -

rp Splash zone correction factor -

rinj Relative number of stress cycles in short term condition (in) for loading condition j -

R Radius of free edge of rounded bracket mm

Ra Surface roughness μm

r Radius of burr grinding disk mm

t Thickness either net tn50 or gross tgr thickness mm

ta Thickness of transverse attachment mm

td Doubling plate thickness mm

teff Effective thickness for butt welds attachments and cruciform joints mm

t1-n50 Net thickness of main plate mm

t2-n50 Net thickness of attachment plate mm

t3-n50 Net thickness of supporting girderframe mm

tp-n50 Plate thickness mm

tf-n50 Flange net thickness mm

tf-gr Flange gross thickness mm

tgr Gross thickness mm

tw-n50 Net web thickness mm

tw-gr Gross web thickness mm

tn50 Net thickness for fatigue assessment with half corrosion addition deducted mm

tn25 Net thickness with 25 corrosion addition deducted mm

tbkt-gr Gross thickness of bracket mm

Sec

tion

1


DNV GL AS

Symbol Meaning Unit

TC Time in corrosive environment during the design fatigue life years

TC25 Time in corrosive environment during the design life years

TD Design life = 25 years years

TDF Design fatigue life to be specified in the design brief years

TDF-C Corrected design fatigue life converting time in corrosive environment to time in air years

TF Predicted fatigue life years

TLC Draft of actual loading condition m

V Vessel speed at 15 sea margin 85 MCR and design draft knots

Vd Vessel contract speed at y sea margin x MCR and design draft knots

W Width of groove mm

xexeFwd

xeAft

Distance from closest end of stiffener span to hot spot Forward and Aft of transversebulkhead

m

xwt External fillet weld length mm

x Longitudinal coordinate in the shiprsquos coordinate system at the load calculation pointpositive forward

m

xs Longitudinal distance from liquid surface centre to load point m

xshift Shifted distance from intersection point to read out point mm

X Stress at 2 000 000 cycles of the FAT class S-N curve Nmm2

y Transverse coordinate in the shiprsquos coordinate system at the load calculation point positiveto port side

m

ys Transverse distance from liquid surface centre to load point m

z Vertical coordinate value in the shiprsquos coordinate system at the load calculation pointrelative to base line positive upwards

m

zs Vertical distance from liquid surface centre to load point m

Zeff-n50ZAft-n50

ZFwd-n50

Net section modulus of stiffener based on effective flange (or hot spot at stiffener web incase of no top stiffener) Aft or forward of transverse bulkhead

cm3

Zn50 Net section modulus of stiffener based on attachment plate with breadth equal to thestiffener spacing

mm

(i) Suffix which denotes dynamic load case HSM FSM BSR-P BSR-S BSP-P BSP-S OST-P orOST-S specified in the rules Pt 3 Ch 4 lsquoi1rsquo denotes dynamic load case (snap shot) HSM-1FSM-1 BSR-1P BSR-1S BSP-1P BSP-1S OST-1P or OST-1S lsquoi2rsquo denotes dynamic loadcase (snap shot) HSM-2 FSM-2 BSR-2P BSR-2S BSP-2P BSP-2S OST-2P or OST-2S

(j) Suffix which denotes loading condition relevant according to the rules

22 AbbreviationsThe abbreviations in Table 2 is used in this CG

Sec

tion

1


DNV GL AS

Table 2 Abbreviation list

FE Finite Element

FEA Finite Element Analysis

FEM Finite Element Model

CG Class Guideline

RP Recommended Practise

CSR Common Structural Rules

COG Centre of gravity

EDW Equivalent Design Wave

LCF Low Cycle Fatigue (implying frequencies from changes in the loading conditions)

HCF High Cycle Fatigue (implying wave frequencies)

H(ωβ) Transfer function

LF Low Frequency (implying slow drift forces or changes to loading conditions)

WF Wave Frequency

HF High Frequency (implying frequencies from vibration)

NA North Atlantic

WW World Wide

PSM Primary Supporting Member eg frames girders stringers and floors

RAO Response Amplitude Operator

TRF Transfer function

STR Short Term Response

LTR Long Term Response

A-E Normal strength (mild) steel

A-E3236 High tensile steel with yield stress of 315355 Nmm2

VBM Vertical Bending Moment

HBM Horizontal Bending Moment

VSF Vertical Shear Force

HSF Horizontal Shear Force

AF Axial Force

NDE Non-Destructive Examination

23 Coordinate system and sign conventionsThe origin of the ldquoright handrdquo coordinate system has co-ordinates (L aft of the foremost end centre linebase line) see Figure 1 x y and z are longitudinal transverse and vertical distance from origin to the loadpoint of the considered structural member A similar coordinate system is given in the rules to define positive

Sec

tion

1


DNV GL AS

motions relevant to define inertia loads or internal pressure loads but the origin is defined at COG insteadPositive global hull girder loads is also defined in the rules and could refer to origin located at the neutralaxis The software may refer to the origin as located at FE minus the rule length Care should be used whenestimating the fatigue stress based on these different definitions

Figure 1 Coordinate system

Tension is defined as positive stress and compression as negative stress eg the hogging wave bendingmoment is defined as positive which gives tension in deck which is consistent with positive z at a hot spotabove the neutral axis

3 Application

31 GeneralThis CG includes procedures for evaluation of fatigue strength based on cyclic stresses from wave loads andship motions for the following but not limited to

mdash steel ship structuresmdash foundations welded to hull structuresmdash any other areas designated as primary structuresmdash connections in way of attachments and outfitting details welded to primary ship structures such as

doubling plates

Sec

tion

1


DNV GL AS

This CG may be adapted for modification of existing ship structures subject to the limitations imposed by theoriginal material and fabrication techniquesThis CG is valid for C-Mn steels duplex and super duplex steels and austenitic steels with yield stress lessthan 500 Nmm2Fatigue capacity for aluminium is not included but reference is given to IIW ref [910]

32 Calculated and observed fatigue strengthReal fatigue performance may differ from design calculations due to influence from important parameters likeworkmanship related to fabrication corrosion protection maintenance and the actual encountered fatigueloading To achieve a good correlation between calculated and the actual fatigue strength the assumptions inthe calculations need to reflect the expected conditions

33 VibrationsThe procedures do not take into account the effect of wave induced hull girder vibrations Guidance on howto take into account the fatigue effect of wave induced vibrations for full body vessels (tankers bulk carriersand ore carriers) is presented in AppI and for container ships in DNVGL-CG-0153 Fatigue and UltimateStrength Assessment of Container Ships Including Whipping and Springing Wave induced vibrations maycontribute to fatigue damage also on other ship typesThe procedures do not include methods for other vibratory effects such as engine or propeller inducedvibrations

34 Low cycle fatigueThe procedures do not take into account the effect of low cycle fatigue Low cycle fatigue eg as occurringduring the loading-unloading cycles or during high utilisation of crane capacity during operation is presentedin AppH

4 Methods for fatigue assessment

41 Stress types used for fatigue assessmentDepending on the stress calculation method and the model representing the ship structure the fatigueassessment may be done for nominal stresses hot spot stresses or local stresses at free plate edge Fatiguelife based on nominal and local stresses can be assessed from nominal stress S-N curves (FAT classes) Hotspot stresses can be calculated based on direct calculations or based on nominal stress times tabulated stressconcentration factors (K) The fatigue assessment in the two latter cases should be done using hotspot stressS-N curves

42 Methods for stress calculationThe long term stress range distribution is fundamental for comparing the fatigue loading with the fatiguecapacity represented by the S-N curve This CG outlines three methods for stress range calculation

1) A rule load assessment of the long-term stress range distribution represented by a dynamic stress rangeat a specific probability level and a straight line spectrum

2) A direct design wave approach calculated for the ship to represent the wave loading and use of a straightline spectrum

3) A spectral method for calculation of the long-term dynamic stress range distribution considering thewave

Sec

tion

1


DNV GL AS

In the first method the loads are based on the equivalent design waves EDW prescribed by the rules andapplied to beam or FE models The load components are combined by prescriptive formulationsIn the second method ship specific design waves are established to represent the loading from the rulesThe design waves may be applied to beam models and various types of finite element models The loadcomponents are combined considering phase relationsIn the third method the long-term stress range distribution is calculated based on directly calculatedhydrodynamic loads for a specific ship and specific wave scatter diagram The loads may be applied tobeam models and various types of finite element models The load components are combined accuratelyconsidering the phase relationsThe different methods are illustrated in Figure 2

Figure 2 Flow diagram over possible fatigue analysis procedures

43 Selection of prescriptive or direct methods for calculation of stressesMethods to be used for the standard ship types are given by the rules

Sec

tion

1


DNV GL AS

Refined analysis may be applied based on

mdash experience with the methods applied to similar structural detailsmdash experience of cracks on similar detailsmdash consequences of a fatigue damage

In general the prescriptive method for fatigue assessment gives a good representation of the fatiguestrength The reliability of the calculated fatigue lives is however assumed to be improved by refined designanalysis

5 Fatigue approaches

51 GeneralDifferent approaches are used depending on efficiency and applicability The following approaches can beapplied for welded joints

mdash Nominal stress S-N curves referred to as FAT classes in Sec2 [2]mdash Hot spot stress S-N curve as described in Sec2 [2]mdash Weld notch stress S-N curve which is not used in this CG reference is made to DNVGL-RP-C203

For base material (free plate edges) the following approach is applicable

mdash Base material S-N curves for local stress at free plate edge depending on the surface conditions asdescribed in Sec2 [2]

Nominal stress Figure 4 can be derived by beam theory using sectional forces and moments or from coarseFE models In a simple plate specimen with an attachment as shown in Figure 3 the nominal stress is themembrane stress An example of fatigue design using this procedure could be the transverse butt welds in adeckHot spot stress Figure 5 is the structural stress which in addition to the nominal stress also contains thepart of the stress increase caused by structural discontinuities and presence of attachments The weld notchstress due to the local weld geometry is excluded from the stress calculation as it is accounted for in thecorresponding hot spot S-N curveThe weld notch stress Figure 5 is defined as the total stress resulting from the geometry of the detail andthe non-linear stress field due to the notch at the weld toe The notch stress approach is also applicableto the weld root The local stress at free plate edge Figure 4 is defined as the stress resulting from thegeometry of the detail at the free plate edge such as a cut-out or hatch corner The local stress needs tobe assessed with finite element analysis for more complex geometries The selection of S-N curve maydepend on the amount and type of inspection during fabrication The size of defects inherent the S-N dataare described in AppGIt should be noted that in welded joints there may be several locations at which fatigue cracks can developeg at the weld toes at the weld roots and in the weld itself Each potential location may need to beconsidered separately

52 Crack failure modesThe following fatigue cracking failure modes are considered see also Figure 3

mdash Weld toe cracking into the base material see Figure 3 a) In welded structures fatigue cracking from weldtoes into the base material is a frequent failure mode The fatigue crack is initiated at small defects orundercuts at the weld toe where the stress is highest due to the weld notch geometry

mdash Weld root cracking through the fillet weld see Figure 3 b) Fatigue cracking from root of fillet welds withcrack growth through the weld is a failure mode that can lead to significant consequences Use of filletwelds should be avoided in highly stressed cruciform joints Non-destructive examination (NDE) is lessreliable in this type of connection compared with a full penetration weld as crack propagation from the

Sec

tion

1


DNV GL AS

root is difficult detect before the crack reach the surface To avoid root failure a sufficient weld throatthickness or partialfull penetration weld is needed see also Figure 2 and Figure 3

mdash Weld root cracking into the section under the weld see Figure 3 c) Fatigue crack growth from the weldroot into the section under the weld has been experienced The number of cycles until failure for thisfailure mode may be of similar magnitude as fatigue failure from the weld toe However it is difficultto improve the fatigue strength except by using alternative types of welds locally eg improvement ofthe weld toe may make the root become the critical location and improvement for the root becomesnecessary This can be obtained by partial or full penetration weld instead of fillet weld close to thetermination of the weld connection

mdash Crack growth from a surface irregularity notch or thermal cutted edge into the base material see Figure3 d) Fatigue cracking in the base material is a failure mode that is relevant in areas with high stressrange Then the fatigue cracks often initiate from notches or grooves in the components or from smallsurface defectsirregularities Examples of such details of concern are cut-outs and hatch corners

a) Fatigue crack growth from the weld toe into the base material

b) Fatigue crack growth from the weld root through the fillet weld

Sec

tion

1


DNV GL AS

c) Fatigue crack growth from the weld root into the section under the weld

d) Fatigue crack growth from a surface irregularity or notch into the base material

Figure 3 Failure modes

53 Assessment using local and nominal stress approachIn cases where nominal stresses are well defined the nominal stress approach can be applied using relatedS-N curves The joint classification and corresponding FAT classes takes into account the local stressconcentrations created by the joints themselves and by the weld profile The design stress is the nominalstress σn which is acting adjacent to the weld under consideration However if the joint is situated in aregion of gross stress concentration Kg resulting from the geometry of the structure this need to be takeninto account as illustrated in Figure 4 and derived as

Sec

tion

1


DNV GL AS

Figure 4 Nominal stress σn and local stress σloc

54 Assessment using hot spot stress approachThe relevant hot spot stress in the parent material at the weld toe is the range of maximum principal stressadjacent to the potential crack location using stress concentration due to the geometry of the structure Asan example for the welded connection shown in Figure 5 the relevant stress is the hot spot stress σHS Forthe weld shown in Figure 3 the stress concentration factors need to include both the local geometry of thetransverse attachment and the holeFor details eg longitudinal end connections or for other details where the stress concentration factordepends on the load component or where there are more than one hot spot it is recommended to use thehot spot stress approach The hot spot stress is then established by the following relation between nominalstress and hot spot stress

where K is the structural stress concentration factor which can be tabulated for standard details and loadcomponents Hot spot stress may also be calculated by use of local FEA

Sec

tion

1


DNV GL AS

Figure 5 Hot spot stress σHS and notch stress σnotch

55 Stress rangeDynamic stress variations are referred to as either stress range (Δσ) or stress amplitude (σ) For linearresponses with zero mean stress the following relation applies

Alternatively if the mean stress differs from zero the stress range can be estimated as

The stress range Ds is being used as the reference in the S-N curves

6 Additional class notations

61 GeneralThe rules RU SHIP Pt 6 specify different class notations This subsection contains an overview of the classnotations which are related to fatigue assessment or damage

62 CSACSA requires that specified structural details should be assessed by directly calculated loads This includestypically longitudinal end connections plates between stiffeners and specified details in the rules For somedetails local FE analysis is required and these local models are sub-models in global FE models

Sec

tion

1


DNV GL AS

63 RSDRSD requires a direct wave load analysis and global FE model for container ships Typical structural detailssubject to fatigue assessment are hatch corners end connections of longitudinals and welded details of upperpart of hull girder

64 PLUSPLUS requires extended assessment for specified structural details based on prescriptive loads in the rules Ittypically includes hot spots at the crossing of longitudinals and web frames (eg slots and lugs) and specifieddetails in the rules The procedure is based on local FE analysis and these local models are sub-models inpartial ship models or global FE models

65 HMONHMON refers to an approved Hull Monitoring System A hull monitoring system including strain sensors isa decision support system (DSS) which gives guidance to the officer on watch on the fatigue (and extreme)loading of the hull structure in general and for specific structural details The measurements capture also thefatigue (and extreme) contribution from wave induced vibrations like springing and whipping The change ofthe fatigue loading due to change of the course and speed are displayed on board and measurement data canfurther be inspected on shore

66 VIBR and COMF-VHigh cycle fatigue damage may also be a result of vibration Hull structure eg frames vibrating due toengine or propeller excitation may give fatigue cracks at odd places Also other equipment may contributeto the excitation For tank structures added mass from the fluid reduces the natural frequencies of thestructural members which then may be excited in resonance The two class notations VIBR and COMF-Vhave the objective of limiting the level of vibration and thereby directly or indirectly will reduce the risk offatigue damage

67 NAUT and routingWeather routing will have a significant impact on wave load level both with respect to extreme loads aswell as fatigue loading There is currently no class notation for weather routing although there are somerequirements to resolution of data from weather information system of weather surveillance system withinthe NAUT notation A HMON system may also be a part of a routing system

68 WIVWave induced vibrations typically designated as whipping and springing cause additional high-frequentdynamic stress in the ship hull structure superimposed to the wave-induced stress The additional stresshas impact on the fatigue loading Class notation WIV can be assigned if additional strength assessment iscarried out with explicit consideration of hull girder vibrations

7 DefinitionsA list of definitions is given in the followingClassified structural detail A structural detail containing a structural discontinuity including a weld or weldsalso referred to as a standard structural detail for which the nominal stress approach is applicable Eachclassified detail is defined to belong to one S-N curve This means that the K-factor for this detail is includedin the S-N curveConstant amplitude loading A type of loading causing a regular stress fluctuation with constant magnitudesof stress maxima and minima

Sec

tion

1


DNV GL AS

Crack growth rate Increase of crack length in m per loading cycleEccentricity Misalignment of plates at welded connections measured in the plate thickness directionEffective notch stress Notch stress calculated for a notch with a certain effective notch radiusFatigue Deterioration of a component caused by crack initiation andor by the growth of cracksFatigue action Load effect causing fatigueFatigue damage ratio For constant amplitude stress it is the ratio of the number of stress cycles to thenumber of cycles to failure from the S-N curve For a stress spectrum it is the sum of the individual fatiguedamage ratio for the different stress range levelsFatigue life For constant amplitude loading it is the number of stress cycles at a particular stress rangerequired to cause fatigue failure of a component For a stress spectrum it is the total number of stress cyclesuntil failureFatigue limit Fatigue strength under constant amplitude loading corresponding to a high number of cycleslarge enough to be considered as infinite by a design code The fatigue limit may be used for screening and itis illustrated in Figure 1Fatigue rate Fatigue damage calculated for a certain time interval divided by the budget damage for thesame time interval The time interval is typical five minutes half an hour or one hour The budget damage isestimated as the ratio of the time interval to the target fatigue life This is used to demonstrate the instantfatigue loading of critical details compared to the design requirementFatigue resistance Structural detailrsquos resistance against fatigue actions in terms of S-N curve or crackpropagation propertiesFatigue strength Magnitude of stress range leading to a particular fatigue life Good fatigue strength thenimplies low stress level or long fatigue lifeFatigue stress Stress including corrections for eg mean stress thickness effect surface finishing andmaterial factor relevant for fatigue assessment purpose ie

mdash Nominal stress for welded detailsmdash Hot spot stress for welded detailsmdash Local stress at free plate edge

Fracture mechanics Assessment of fatigue by crack propagation analysis until unstable fracture occursGeometric stress See ldquohot spot stressrdquoHot spot Hot spots are locations in the structure where fatigue cracks may initiate due to the combinedeffect of nominal structural stress fluctuation and stress raising effects due to the weld geometryHot spots may be located at

mdash Weld toesmdash Weld root of partial penetration or fillet welds

Hot spot stress The value of stress on the surface at the hot spot (also known as geometric stressor structural stress) Hot spot stress is eg the stress at the weld toe taking into account the stressconcentration due to structural discontinuities and presence of welded attachments but disregarding thenon-linear stress distribution caused by the notch at the weld toe The hot spot stresses to be consideredcorrespond to the two principal stresses on the plate surface at the weld toe The first principal stress actswithin plusmn45deg perpendicular to the weld and the second principal stress acts outside plusmn45deg The hot spot stressshould be obtained by multiplying the nominal stress by tabulated Stress Concentration Factor (K) or directlyby FE hot spot analysisLocal notch A notch such as the local geometry of the weld toe including the toe radius and the anglebetween the base plate surface and weld reinforcement The local notch does not alter the structural stressbut generates non-linear stress peaksLocal stress at free edge Stress at a free plate edge derived using finite element analysis or tabulated StressConcentration Factors (K)

Sec

tion

1


DNV GL AS

Macro-geometric discontinuity A global discontinuity the effect of which is usually not taken into accountin classified structural details such as large openings a curved part in a beam a bend in a flange andeccentricity in lap jointsMacro-geometric effect A stress raising effect due to macro-geometry in the vicinity of the welded joint butnot due to the welded joint itselfMembrane stress Average normal stress across the thickness of a plateMiner sum Summation of individual fatigue damage ratios caused by each stress cycle or stress range blockaccording to the Palmgren-Miner ruleMisalignment Axial and angular misalignments caused either by detail design or by fabricationNominal stress Nominal stress is the stress in a structural component taking into account macro-geometriceffect including effective breadth of flanges but disregarding the stress concentration due to structuraldiscontinuities and the presence of welds Nominal stress should be obtained either using coarse or fine meshFE analysis or using analytical calculation based on beam theory with effective breadth of flanges includedNon-linear stress peak The stress component of a notch stress which exceeds the linearly distributedstructural stress at a local notchNotch stress Total stress of a notch at a free plate edge weld root or weld toe taking into account thestress concentration caused by the local notch Thus for welds the notch stress consists of the sum ofstructural stress and the non-linear stress peakNotch stress concentration factor The ratio of notch stress to the nominal stressParisrsquo law An experimentally determined relation between crack growth rate and stress intensity factor range(Fracture mechanics)Palmgren-Miner rule Linear damage accumulation for variable amplitude loading using S-N curves Fatiguefailure is expected when the Miner sum reaches unityRainflow counting A standardised procedure for stress range countingShell bending stress Bending stress in a shell or plate-like part of a component linearly distributed acrossthe thickness as assumed in the theory of shellsS-N curve Graphical presentation of the dependence of fatigue life (N) on fatigue strength (S)Stress cycle A part of a stress history containing a stress maximum and a stress minimumSCF Stress concentration factorStress intensity factor Factor used in fracture mechanics to characterise the stress at the vicinity of a cracktipStress range The difference between stress maximum and stress minimum in a stress cycleStress range block A part of a total spectrum of stress ranges which is discretised in a certain number ofblocksStress range exceedance A tabular or graphical presentation of the cumulative frequency of stress rangeexceedance ie the number of ranges exceeding a particular magnitude of stress range in stress historyHere frequency is the number of occurrencesStress ratio Ratio of minimum to maximum value of the stress in a cycleStructural discontinuity A geometric discontinuity due to the type of welded joint usually found in tables ofclassified structural details The effects of a structural discontinuity are (i) concentration of the membranestress and (ii) formation of secondary bending stressStructural stress A stress in a component resolved taking into account the effects of a structuraldiscontinuity and consisting of membrane and shell bending stress components Also referred to asgeometric stress or hot spot stressStructural stress concentration factor The ratio of hot spot (structural) stress to nominal stress alsoreferred to as geometrical stress concentration factor In this classification note the shorter notation ldquoStressconcentration factor due to geometryrdquo (K) is usedVariable amplitude loading A type of loading causing irregular stress fluctuation with stress ranges (andamplitudes) of variable magnitude

Sec

tion

1


DNV GL AS

Weld shape factor The ratio between the notch stress and the structural (hot spot) stress describing theeffect of the weld contour This is part of the notch stress concentration factor The weld shape factor Kw isincluded in the hot spot S-N curve

Sec

tion

2


DNV GL AS

SECTION 2 FATIGUE CAPACITY

1 IntroductionThe main principles for fatigue capacity are described in this section The basis is the design S-N curves in[2] but the following should be considered in assessment of the fatigue strength

mdash Corrosive environment or in air environment [3]mdash Mean stress effect [4]mdash Thickness effect [5]mdash Material effect [6]mdash Post-weld treatment Sec7

Additional stresses resulting from fabrication tolerances of butt welds and cruciform joints should beconsidered when the fabrication tolerances exceed that inherent in the S-N data Reference is made to AppA

2 S-N curves

21 GeneralThe fatigue capacity of welded joints and base material is defined by S-N curves which are obtained fromfatigue tests The design S-N curves are based on the mean-minus-two-standard-deviation curves which areassociated with a 975 probability of survival A failure refers to a through thickness crack with reference toAppBThe S-N curves are applicable for normal and high strength steels used in construction of hull structures upto 500 Nmm2 Other S-N curves are given in AppBThe S-N curves for welded joints include the effect of the local weld notch for the hot spot approach Thisalso means that if a butt weld is machined or ground flush without weld overfill a better S-N curve can beused For the nominal stress approach ie FAT class S-N curves also the detail itself is included

22 S-N curves and detail classificationFor practical fatigue design welded details and base material can be divided into several classes each with acorresponding design S-N curve Typical details may fall in one of the classes specified in Table 1 and Table 2depending upon

mdash The geometry of the detailmdash The direction of the fluctuating stress relative to the detailmdash The method of fabrication (misalignmentdefects and surface condition) and inspection of the detail

Each detail at which fatigue cracks may potentially develop should where possible be classified inaccordance with tables given in AppA for welded joints and in accordance with Table 2 for base materialEach potential crack location should be classified separatelyThe basic design S-N curve is given as

with S-N curve parameters given in Table 1 and where

N = Predicted number of cycles to failure for stress range Δσ

Δσ = Stress range in Nmm2

Sec

tion

2


DNV GL AS

m = Negative inverse slope of S-N curve

log K2

= Intercept of log N-axis by S-N curve

where

K1 = Constant of mean S-N curve (50 probability of survival)

K2 = Constant of design S-N curve (975 probability of survival)

δ = Standard deviation of log N

= 020

23 S-N curves for in-air environmentS-N curves for in-air environment are given in Table 1 and in Figure 2 up to 108 cyclesThe reference stress ranges at 2middot106 and 107 cycles are included and also the structural stress concentrationfactor as derived from the hot spot method is included for reference For welded details the D-curve is thereference curve for the hot spot stress method For base material the B1 curve is used as the referenceThe B B2 C C1 and C2 curves are applicable for local stress at free plate edges according to Table 2 Theparameterized S-N curve designated with FAT X (where X is the stress range at 2middot106 cycles see Figure 1) isused for welded details assessed with the nominal stress approach and should not be used in relation to thefree plate edges

Figure 1 Illustration of FAT class S-N curves with FAT X number

The B1 curve forms the upper limit of all other curves at very low number of cycles ie for well below 10000cycles This is regarded relevant for certain details for low cycle fatigue

Sec

tion

2


DNV GL AS

Table 1 S-N parameters for air

S-NCurve

Referencestress at 2middot106

cycles (FAT)

Reference stressat 107 cycles

(knuckle) Δσq

Structural stressconcentrationembedded in the detailand taken at 2middot106

(107) cycles

N le 107 N gt 107

Nmm2 Nmm2 - log K2 m log K2 m+Δm

B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5

FAT X 0585middotX 90X (90X) 6301+3middotlog(X) 3 7+5middotlog(0585middotX) 5

Figure 2 S-N curves in air

Sec

tion

2


DNV GL AS

Table 2 S-N-curve depending on surface finishing for base material and free plate edge

Joint configuration fatigue cracklocation and stress direction Description of joint S-N curve

FAT classNmm2

1

Base material Rolled or extruded plates andsections as well as seamless pipes No surfaceor rolling defects

B1

160

2

Free plate edge Machine cutting eg by athermal process or shear edge cutting Smoothsurface free of cracks and notches Cuttingedges chamfered or rounded by means ofsmooth grinding groove direction parallelto the loading direction Stress increase dueto geometry of cut-outs to be considered bymeans of direct numerical calculation of theappertaining maximum local stress range

B

150

3

Free plate edge Machine cutting eg by athermal process or shear edge cutting Smoothsurface free of cracks and notches Cuttingedges broken or rounded

B2 140

4

Free plate edge Machine cutting eg by athermal process or shear edge cutting No edgetreatments Surface free of cracks and severenotches

C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5

Free plate edge Manually thermally cut eg byflame cutting With edge treatments Surfacefree of cracks and severe notches

C1 112

6

Free plate edge Manually thermally cut eg byflame cutting No edge treatments Surface freeof cracks and severe notches

C2

100

Note 1 Stress increase due to geometry of cut-outs to be considered

24 S-N curves for stress along the weldsThe S-N curves given in Table 1 are developed for principal stresses acting normal to the weld and should beused together with the maximum stress range within plusmn 45ordm of the normal to the weld as explained in Sec6[13]If the governing stress direction is parallel with the weld direction a stress reduction factor KP should beused on the principal stress range before entering stress into the S-N curve D The stress reduction factorwill depend on the quality of the weld Table 3 Kp is by default 10 for stress direction normal to the weldInstead of using stress reduction factor Kp for the hot spot stress approach also alternative S-N curves canbe applied as given in Table 3 Alternatively the procedure of effective hot spot stress described in Sec6[13] and AppE may be usedThe correction needs to be applied for the nominal stress approach if a parametrized FAT X class is chosenwhich is applicable to a governing stress direction parallel to the weld direction

Sec

tion

2


DNV GL AS

Table 3 Stress reduction factor Kp and FAT classes for longitudinally loaded welds

KpFATNmm2

Figure Description Requirement

072125

1Automatic welds carriedout from both sides

1No start-stop position is permittedexcept when the repair is performedby a specialist and inspection carriedout to verify the proper execution ofthe repair

080112

2 Automatic fillet orbutt welds carried outfrom both sides butcontaining stop-startpositions3 Automatic butt weldsmade from one sideonly with a backingbar but without start-stop positions

3 When the detail containsstart-stoppositions use Kp = 090 or FAT 100

090100

4 Manual fillet or buttwelds5 Manual or automaticbutt welds carried outfrom one side onlyparticularly for boxgirders

6 Repaired automaticor manual fillet or buttwelds

5 A very good fit between the flangeand web plates is essential Preparethe web edge such that the root faceis adequate for the achievement ofregular root penetration with outbrake-out

6 Improvement methods that areadequately verified may restore theoriginal category

Sec

tion

2


DNV GL AS

KpFATNmm2


11380

Intermittentlongitudinal fillet weld(based on stress rangein the flange at weldends)

In presence of shear stress τ theFAT class and Kp factor should bedowngraded by the factor 1-ΔτΔσbut not below FAT 36 or Kp of 25

25 Nominal stress S-N curvesFor welded details of which the fatigue capacity does not strongly depend on the type of loading ie axial orbending the nominal stress approach can be used based on S-N curves from Table 1 designated as FAT X InAppA the FAT classes are included in parallel with stress concentration factors K for various detailsFAT class curves are not given for the structural stress concentration factors referring to typical endconnections since for these there are separate geometrical stress concentration factor for axial loading Kaand bending Kb

26 Other steel typesFor Duplex Super Duplex and austenitic steel one may use the same S-N curve as for C-Mn steels

3 Corrosive environment

31 GeneralFatigue life of steel structures in freely corroding condition eg submerged in sea water is shorter than in-air environment For coated or cathodically protected steel in sea water approximately the same fatigue lifeas in dry air is obtainedThe fraction of time in-air and in corrosive environments depends on the location of the detail and is specifiedby the rules For details being partly in air and partly in corrosive environment the fatigue damage may becalculated according to [3]The general principle is that all details should be protected from freely corroding conditions which meansthat the dominating part of the design fatigue life is associated with the in-air environment while a minorpart is considered as corrosive

32 Fatigue assessment in corrosive environmentFor details in corrosive environment like sea water the S-N curves for in-air environment can be used butthe fatigue damage in corrosive environment should be multiplied with a factor of 20 The same effect isachieved if the increase of the fatigue design life is assumed as described in Sec3 [3]

4 Mean stress effectFor fatigue analysis the stress range may be reduced dependent on whether the cycling stress is in tension orin compression

Sec

tion

2


DNV GL AS

The basis for the mean stress effect is the variable amplitude stress range Δσ at a 10-2 probability level ofexceedance for the specific wave environment This implies that Δσ should include the environmental factorfe when the prescriptive loads are used

41 Base material and free plate edgesMean stress σmean refers to the static stress If the stress concentration factor is included it needs to beincluded consistently in the mean stress and the dynamic stress range The calculated stress range should bemultiplied with the mean stress reduction factor fmean illustrated in Figure 3 and given as

Figure 3 The stress range reduction factor for base material and free plate edges

42 Welded jointsThe mean stress reduction factor fmean for welded details is illustrated in Figure 4 and given as

Sec

tion

2


DNV GL AS

Figure 4 The stress range reduction factor for welded details

5 Thickness effect

51 GeneralThe fatigue strength of welded joints depends to some extent on the plate thickness This effect is due to thelocal geometry of the weld toe in relation to thickness of the adjoining plates eg weld profiling may affectthe thickness effect as given in AppF It is also dependent on the stress gradient over the thickness ie itcan differ between axial and bending loadingThe correction factor fthick of the stress range is given as

for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where

n = Thickness exponent

teff = Effective thickness according to [52] If nothing else is specified tn50 should be used

The thickness exponent n on the fatigue strength is given in Table 4 for specific details If nothing else isstated an exponent of 020 may be used for welded details

Table 4 Welded details thickness exponents

No Detail categorydescription

Geometry Condition n

As-welded 021 Cruciform jointstransverse T-joints plateswith transverseattachments

Weld toe treated bypost-weld improvementmethod

02

As-welded 022 Transverse buttwelds

Ground flush or weldtoe treated by post-weldimprovement method

01

Sec

tion

2


DNV GL AS



Any 013 Longitudinal welds orattachments to plateedges


01

Any 004 Longitudinalattachments onthe flat bar or bulbprofile

Weld toe treated bypost-weld improvementmethod 1)

00

Sec

tion

2


DNV GL AS



As-weldedt2 ge 033 t1else

0100

5 Longitudinalattachmentsunsupportedand supportedlongitudinally


00

As-welded 016 Doubling platesunsupported andsupported

Weld toe treated by post-weld treatment method

00

1) No benefit applicable for post-weld treatment of longitudinal end connections

52 Effective thickness of weldsThe thickness exponent is considered to account for the size of the plate through which a crack will mostlikely grow To some extent it also accounts for local geometry at the weld toe However the equation in[51] does not account for the weld width

Sec

tion

2


DNV GL AS

The thickness effect depends also on the weld width d of butt welds and attachment length of cruciformconnections as illustrated in Figure 5 For butt welds where d may differ on the two plate sides the width dfor the considered hot spot side should be used The thickness effect is lower for short lengths of d Thus thethickness t for butt welds and cruciform joints should be replaced by an effective thickness in mm which canbe derived as

where the parameters d and tn50 are measured in mm and are defined in Figure 5

Figure 5 Definition of attachment length of cruciform joints and weld width for butt welds

If the details of the weld are not shown on the structural drawings an estimate of the width in mm can betaken as

53 Thickness effect of welds in prescriptive analysis of longitudinal endconnectionsFor prescriptive stress analysis the following effective thickness should be considered

mdash Flat bar and bulb profile No thickness effectmdash Angle bar and T-bar Flange thickness teff = tf-n50 and a thickness exponent n = 01 when the hot spot is

located on the stiffener flange

54 Thickness effect of welds in FE analysisFor FE analysis the effective thickness to be considered is the thickness of the member where the crack islikely to initiate and propagate

Sec

tion

2


DNV GL AS

55 Thickness effect of base materialThe thickness effect is not regarded applicable for base material and free plate edges assuming goodworkmanship practise (S-N curve B1 B B2 and C) and when adequate protective measures are takenagainst wear tear and corrosion eg from friction from cargo or cargo handling appliancesThe thickness exponent of n = 01 can be assumed if the above is not fulfilled (S-N curve C1 and C2)

6 Material factorThe fatigue stress range Δσ should include a correction factor for the material strength fmaterial taken as

where the specified minimum yield stress ReH is in Nmm2

Sec

tion

3


DNV GL AS

SECTION 3 FATIGUE STRENGTH REPRESENTATION

1 IntroductionThe basis for the fatigue assessment is an estimate of the stress range (hot spot nominal or local at freeplate edge) which is described in [2] When assessing the fatigue strength the time in corrosive and in-air environment the time in different loading conditions and the time in port need to be accounted for asdescribed in [3] The results of the fatigue assessment can be presented in different ways depending on whatis found convenient

mdash Calculated fatigue damage versus allowable fatigue damage [3]mdash Calculated fatigue life versus design fatigue life [3]mdash Actual stress range (calculated) versus permissible peak stress range (calculated) [4]mdash Given stress concentration factor versus permissible stress concentration factor (calculated) [5]mdash Given FAT class versus required FAT class (calculated) [5]

The scope and extent of the fatigue assessment the loading conditions and related parameters to considerand the time at sea are given in the rules for the standard ship types

2 Fatigue stress range

21 GeneralThe calculated stress range should follow the net scantling approach tn50 for prescriptive fatigue assessmentand net or gross t based on FE analysis where net or gross scantlings is defined in the rules for the specificship types Since there is a difference on how the corrosion affects the global and local stress a correctionfactor fc needs to be included In addition the calculated stress depends on the wave environment which isrepresented by the environmental factor feBefore performing fatigue strength assessment the calculated stress entered into the S-N curve needs to becorrected for other fatigue capacity effects such as

mdash Mean stressmdash Thicknessmdash Materialmdash Post-weld treatment

Results from performed fatigue analyses are presented in AppC in terms of allowable stress ranges asfunction of the Weibull shape parameter The basis for the allowable stress ranges is that long term stressranges can be described by a two parameter Weibull distribution

22 Scantlings approach factor fcThe global stress is less affected by the corrosion than the local stress due to the probabilistic nature ofcorrosion which gives less average corrosion on a global level compared to at a local level Thus the localstress should be based on tn50 while the global stress should be based on tn25For prescriptive fatigue assessment using a beam model based on tn50 the correction factor fc only applies tothe global stresses When using a global FE model based on tn50 or tgr the same correction factor applies toboth the global and local stresses ie directly on the stressThe correction factor should be taken as

a) Prescriptive fatigue strength assessment

mdash fc = 10 for local plate and stiffener bending based on tn50

mdash fc = 095 for hull girder stresses based on tn50

Sec

tion

3


DNV GL AS

b) FE based fatigue strength assessment

mdash fc = 10 for stresses based on tgr

mdash fc = 095 for stresses based on tn50

Since FE models may be based on either tn50 or tgr the notation t is used for the thickness in relation toFE analysis However prescriptive assessment should be based on tn50

23 Environmental factor feThe environmental factor fe is a correction factor representing the ratio of the stress range based on thevertical wave bending moment from a specific trade and the North Atlantic wave environmentFor prescriptive loads the following environmental factors are used

mdash fe = 10 for North Atlantic wave environmentmdash fe = 08 for World Wide wave environment

When the stress range is calculated based on direct hydrodynamic analysis for a specific trade (or site)through a component stochastic or full stochastic analysis the fe factor is 10 (as long as the waveenvironment is according to the requirements in the rules)

24 Post-weld treatmentReference is made to the rules for the application of the post-weld treatment There are several limitationsfor the benefit of post-weld treatment This factor should preferably not be considered in the initial designphase For burr grinding the stress range can be multiplied with the following factor

where

fT = Fatigue life factor as given in Sec7 [23]

25 Fatigue stress rangeThe fatigue stress range ΔσFS in Nmm2 to be used in the assessment is based on the calculated stress rangeΔσ times the corrections for mean stress effect thickness effect material factor scantlings approach factorthe environmental factor and the post-weld treatment factor

The stress range should be estimated at a 10-2 probability level of exceedance

The straight line spectrum ie 2-parameter Weibull distribution with Weibull slope ξ = 10 is displayedin Figure 1 where it can also be seen that in this case the permissible stress range at a probability level of10-2 is 25 of the stress range at a probability level of 10-8 In Figure 2 the distribution of the damage isillustrated at different probability levels based on the straight line spectrum suggesting that most of thefatigue damage comes from smaller stress ranges close to a probability level of 10-2

Sec

tion

3


DNV GL AS

Figure 1 Straight line spectrum giving a damage of 10 based on S-N curve DFAT 90

Figure 2 Damage distribution for a straight line spectrum based on S-N curve DFAT 90

3 Fatigue damage and fatigue life calculation

31 Fatigue damage accumulation with Palmgren - Minerrsquos ruleFatigue assessment may be carried out by estimating the linear cumulative fatigue damage D by usingPalmgren-Minerrsquos rule formulated as

Sec

tion

3


DNV GL AS

where

ni = Number of cycles at stress range ΔσiNi = Number of cycles to failure at stress range Δσintot = Total number of stress range blocksi = Stress range block index

The number of load cycles ni is determined by the long term stress distribution eg represented by theWeibull distribution while the number of cycles to failure Ni is represented by the S-N curve

The acceptance criteria to the damage D is given by the rules

32 Time in air and corrosive environmentThe design fatigue life is divided into two time periods due to limitation of the corrosion protection It isassumed that the corrosion protection (ie coating system) is only effective for a limited number of yearsduring which the structural details are exposed to in-air environment During the remaining part of thedesign life TC as specified in the rules the structural details are unprotected ie exposed to corrosiveenvironment

33 Annual fatigue damageThe annual fatigue damage is the damage in-air accumulated during a specific loading condition (j) assumed100 time at sea The annual fatigue damage can be based on the fatigue stress range obtained from thepredominant load case (EDW giving the highest and thereby the most representative fatigue stress rangeΔσFS) in the prescriptive fatigue assessment This can be calculated based on a closed form formulation as

Sec

tion

3


DNV GL AS

where

ND = Total number of stress cycles possible per year where 1(4logL) is assumed as the zero up-crossingfrequency

ΔσFS(j) = Fatigue stress range from the predominate load case at the reference probability level ofexceedance of 10-2 in Nmm2

NR = Number of cycles corresponding to the reference probability of exceedance of 10-2 NR = 100

ξ = Weibull shape parameter ξ = 1

Γ(ξ) = Complete Gamma function

K2 = Constant of the design S-N curve as given in Table 1 for in-air environment

μ(j) = Coefficient taking into account the change of inverse slope of the S-N curve m

γ(aξ) = Incomplete Gamma function

Δσq = Stress range in Nmm2 corresponding to the intersection of the two segments of design S-N curveat N = 107 cycles as given in Table 1

Δm = Change in inverse slope of S-N curve at N=107 cycles Δm = 2

34 The combined fatigue damageThe combined fatigue damage is the damage accumulated in both in-air and in corrosive environment for aspecific loading condition (j) during a specified design fatigue life TDF with 100 time at sea The combinedfatigue damage for the design fatigue life TDF for each loading condition (j) can be calculated as

where

TC25 = Time in corrosive environment in years within the duration of the design life TD

TD = Design life in years taken as TD = 25 years

DE(j) = The annual fatigue damage for in-air environment for loading condition (j)

35 The total fatigue damage from multiple loading conditionsTotal fatigue damage within the design fatigue life TDF is the sum of the combined fatigue damages obtainedfor all loading conditions and accounting for the fraction of the time at sea and the fraction in each loadingcondition The total fatigue damage for all applicable loading conditions is calculated as

Sec

tion

3


DNV GL AS

where

f0 = Factor taking into account time in seagoing operations given in the rules

αj = Fraction of time in each loading condition given in the rules

nLC = Total number of loading conditions given in the rules

D(j) = Combined fatigue damage for each applicable loading condition

36 Fatigue lifeThe predicted fatigue life TF is based on different part times in-air and corrosive environment For shortfatigue lives less than TD-TC25 years the predicted fatigue life is based on in-air environment For predictedfatigue lives between TD and TC25 the time after TD-TC25 is based on corrosive environment For predictedfatigue lives above TD it is assumed that the predicted fatigue life is based on a regular maintenanceintervals from the delivery of the vessel This may be approximated by a constant annual damage rate whichalso reflects a relative amount of corrosive environment The predicted fatigue life above TD = 25 years isindependent on design fatigue life TDF (which is equal or longer than TD) If the design fatigue life is changedat a later time due to the need for life time extension the predicted fatigue life for the same operation areadoes not change The formulation is illustrated in Figure 3 and the predicted fatigue life is calculated as

for

for

for

where

and where

D = Total damage given by [35]

Sec

tion

3


DNV GL AS


Figure 3 Illustration of fatigue damage accumulation for different assumptions of time incorrosive environment

4 Permissible stress rangeThe stress range should be less or equal to the permissible stress range

In general the permissible stress range Δσperm can be assessed iteratively as the stress range which yieldsa total damage D=1 However the assessment of Δσperm is limited to cases where there is only one loadingcondition applicable for the whole design fatigue life However it can be used on several loading conditionsto see which is worst The permissible stress range at 10-2 probability level can be calculated as

FAT = Reference stress in Nmm2 range of the S-N curve at 2middot106 load cycles (FAT class) in Nmm2

fN ==

=

=

Factor for the spectrumrsquos shape (straight-line) and the number of load cycles0786 log(NDmiddotf0middotTDF-C)-4966 forwelded details with S-N curves D E F and FAT according to Sec2 Table 10681 log(NDmiddotf0middotTDF-C)-4188 for base material curve B1 and free plate edges with curves B and B2 Sec2Table 1

06007 log(NDmiddotf0middotTDF-C)-3578 for free plated edges with S-N curves C C1 and C2 Sec2 Table 1

Sec

tion

3


DNV GL AS

To reflect the part time in the different environments (protected or corrosive) for the determination of fN thedesign fatigue life should be taken as

5 Permissible stress concentration factor and required FAT classIn general the permissible stress concentration factor or required FAT class can be estimated iterativelyyielding a total damage D=1 In cases where there is only one loading condition applicable for the wholedesign fatigue life the permissible stress concentration factor Kperm can be estimated as

where

Δσn = Nominal stress range of a straight-line spectrum at 102 load cycles in Nmm2

The required reference stress range of the S-N curve at 2middot106 load cycles FATreq can be estimated as

for nominal stress where

Δσ = Nominal stress range or local stress range at free plate edge of a straight-line spectrum at 102 load cyclesin Nmm2

The estimation of Kperm or FATreq should be limited to cases where all load components refer to the samestress concentration factor As the stress concentration factors K for bending and axial loading of longitudinalend connections may differ this method is not regarded applicable or too conservative for these details Ifdifferent loading conditions are to be assessed for fractions of the time of the design fatigue life Kperm orFATreq can only be estimated iteratively yielding a total damage D=1

Sec

tion

4


DNV GL AS

SECTION 4 PRESCRIPTIVE FATIGUE STRENGTH ASSESSMENT

1 Introduction

11 Stress approach for longitudinal end connectionsThis section outlines the prescriptive procedure to determine the stress ranges to be used in the fatigue lifepredictions of longitudinal end connections The approach is based on the hot spot stress approach usingequivalent design waves (EDW)The stress ranges and mean stresses in way of each longitudinal end connection as shown in Figure 3should be evaluated at the flange (for most details) at the following locations

mdash At transverse web frames where relative displacement can be neglectedmdash At transverse bulkheads (swash bulkheads or stools) where relative displacement should be considered

Stress concentration factors for unsymmetrical stiffener geometry and stiffener end connection geometry attwo potential hot spots (point lsquoArsquo and lsquoBrsquo) should be applied according AppAThe hot spot stress approach uses the following

mdash Nominal stresses from beam theorymdash Stress concentration factors K for the different longitudinal end connections as specified in AppAmdash Loading conditions as specified in the Rules

12 Calculation procedure for longitudinal end connectionsAssessment of the fatigue strength of structural members includes at least the following seven steps

a) Selecting type of longitudinal end connectionb) Calculation of wave induced loads from equivalent design waves (EDW)c) Calculation of stress ranges for global and local stress components for each EDWd) Combination of global and local stress components for each EDWe) Selection of predominate EDW ie the EDW giving the highest stress rangef) Use of the S-N curve Dg) Calculation of the cumulative damage or the fatigue life

A flow chart of the calculation procedure is given in Figure 1 and includes also the case when other detailsthan longitudinal end connections are covered by FE models For description of FE analysis reference is madeto Sec6

Sec

tion

4


DNV GL AS

Figure 1 Flow diagram for prescriptive fatigue calculations

13 Definition of stress componentsA schematic illustration of different stress components is given in Figure 2 It shows four levels of stresscomponents which may be relevant ie from global hull girder loads bending of primary supportingmembers (PSM) local stiffener bending and plate bending

Sec

tion

4


DNV GL AS

Figure 2 Definition of stress components

The dynamic global hull girder stress amplitude in Nmm2 is denoted

σG = wave induced hull girder stress amplitude from global loads as vertical bending horizontal bending torsion(warping) and axial force (the latter is neglected in prescriptive assessment)

The dynamic stress amplitude due to bending of PMS is denoted

σdh = bending stress amplitude of PSM eg double hull or framegirder grillage system due to dynamic externalandor internal pressure loads on any side

The dynamic local bending stress amplitudes are denoted

σL = local stiffener bending stress amplitude between two framesbulkheads due to pressuresloads on any side

σd = local relative deflection stress between two framesbulkheads from bending of PSM due to pressureloadson any side

σP = local plate bending stress between frames or stiffeners due to pressureloads on any side

14 Calculation of stress componentsThe stress response in stiffeners and plating subjected to axial loading due to hull girder bending and localbending due to lateral pressures can be calculated based on beam theory combined with tabulated valuesof stress concentration factors Tabulated stress concentration factors are listed in AppA For details witha more complex stress response andor where tabulated values of stress concentration factors are notavailable the stress response should be calculated by FE analyses as described in AppEFor the nominal stress approach instead of stress concentration factors detail specific FAT classes need to beapplied

Sec

tion

4


DNV GL AS

The stress formulas may be combined with prescribed loads derived according to the rules or serve as basisfor determination of stress component factors for a component stochastic fatigue analysis as described inSec5

2 Wave induced loadsSpecification of the prescriptive wave induced loads for vertical and horizontal bending torsion (warping)external and internal pressures are given in the rules The prescriptive wave induced loads are based on theEquivalent Design Wave (EDW) approach with loads referring to a 10-2 probability level of exceedanceThe prescriptive wave induced loads are used in the prescriptive fatigue analysis based on beam theory butthe EDWs may also be used for FE analysis of cargo hold partial ship or global FE models


31 Hot spot stress or nominal stressThe equations in [32] [33] and [34] reflect the hot spot stress approach This is however also applicablefor the nominal stress approach In the latter case the stress concentration factors Ka (for axial stress) andKb (for bending stress) in the formulas in [4] [5] [6] and [7] need to be omitted Further the calculatedstress ΔσHS and ΔσBS in [32] is replaced by the nominal stress range Δσn and local stress range at free plateedge Δσloc respectively

32 Prescriptive fatigue stress rangeThe fatigue stress range for each load case of each loading condition needs to be determined The stressrange of each loading condition (j) to be considered is the stress range obtained from the predominant loadcase according to the rules The stress to be considered for the fatigue assessment is

where

ΔσFSi(j) = Fatigue stress range in Nmmsup2 for load case (i) of loading condition (j)

The fatigue stress range ΔσFSi(j) in Nmm2 corrected for mean stress effect thickness effect materialfactor post-weld treatment and environmental factor is taken as

where

ΔσFSi(j) = Calculated fatigue stress range for load case i in loading condition (j) in Nmm2

ΔσHSi(j) = Hot spot stress range for load case i in loading condition (j) in Nmm2

ΔσBSi(j) = Local stress range at free plate edge for load case i in loading condition (j) in Nmm2

fmean i(j) = Correction factor for mean stress for load case i in loading condition (j)

fe = Environmental factor ref Sec3 [23]

Sec

tion

4


DNV GL AS

fthick = Thickness effect ref Sec2 [5]

fmaterial = Material factor ref Sec2 [6]

fw = Post-weld treatment factor ref Sec3 [24]

The correction factor for the mean stress effect for welded joints is calculated as

where

σmeani(j) = fatigue mean stress for the welded joints for load case i in loading condition (j) in Nmm2

For the base material free plate edge then the mean stress effect is estimated as

33 Stress rangeThe stress range in Nmm2 due to dynamic (wave induced) loads for load case (i) of loading condition (j) isobtained from the following formula

where

σGDi1(j) σGDi2(j) = Stresses due to wave induced hull girder loads in Nmmsup2 as defined in [41]

σLDi1(j) σLDi2(j)= Stresses due to wave induced local bending in Nmmsup2 as defined in [61]

σdDi1(j) σdDi2(j) = Stresses due to wave induced relative displacement in Nmmsup2 as defined in [77]

σdhDi1(j) σdhDi2(j) = Stresses due to wave induced bending of PSM in Nmm2 as defined in [51]

34 Mean stressThe mean stress in Nmm2 due to static and dynamic loads for load case (i) of loading condition (j) isobtained from the following formula

Sec

tion

4


DNV GL AS

where

σGS (j) = Stress due to still water hull girder bending moment in Nmmsup2 as defined in [43]

σLS (j) = Stress due to still water local bending in Nmmsup2 as defined in [62]

σdS (j) = Stress due to still water relative displacement in Nmmsup2 as defined in [43]

σdhS (j) = Stress due to still water bending of PSM in Nmmsup2 as defined in [52]

σmLDi(j) = Mean stress contribution due to wave induced local bending in Nmmsup2 as defined above

σLDi1(j) σLDi2(j) = Stresses due to wave induced local bending in Nmmsup2 as defined in [61]

σmGDi(j) = Mean stress contribution due to wave induced hull girder loads in Nmmsup2 as defined above


σmdDi(j) = Mean stress contribution due to relative deflection in Nmmsup2 as defined above

σdDi1(j) σdDi2(j) = Stresses due to wave induced relative deflection in Nmmsup2 as defined in [75]

σmdhDi(j) = Mean stress contribution due to bending of PSM in Nmmsup2 as defined above

σdhDi1(j) σdhDi2(j) = Stresses due to wave induced bending of PSM in Nmmsup2 as defined in [5]

4 Global hull girder stress

41 Stress due to wave induced hull girder loadsThe hull girder hot spot stress (or nominal stress with Ka=1) in Nmm2 for maximum and minimum loadcomponent i1 and i2 (load case i) of loading condition (j) is obtained from the following formula

Sec

tion

4


DNV GL AS

where

Mwv-LC ik = Vertical wave bending moment in kNm for load case i in loading condition (j) for maximum andminimum load component k (ie snap shot 1 or 2)

Mwh-LC ik = Horizontal wave bending moment in kNm for load case i in loading condition (j) for maximum andminimum load component k

Mσ-wt-LC ik = Dynamic bi-moment related to warping stresses in kNm2 for load case i in loading condition (j) formaximum and minimum load component k

Iω-n50 = Net sectional moment of inertia in m6 of the hull transverse cross section

ωi-n50 = Net sectorial coordinate in m2 of the point being considered in the transverse cross section

Mσ-wt-LC ik is only relevant for vessels with large deck openings according to definition in the rules and mayapply to container vessels general cargo vessels multi purpose vessels and bulk carriers

42 Hull girder vibrationsIn addition to the wave induced hull girder stresses the waves induced hull girder vibrations cause additionalhull girder stresses Guidance on the how to account for the effect of wave induced vibration for blunt ships(eg tankers bulk carriers and ore carriers) is given in AppI and for container vessels in DNVGL-CG-0153Fatigue and ultimate strength assessment of container ships including the effect of whipping and springing

43 Stress due to still water hull girder bending momentThe hull girder hot spot stress or nominal stress (with Ka=1) due to still water bending moment in Nmm2in loading condition (j) is obtained from the following formula

where

Msw = Permissible still water vertical bending moment in kNm

β(j) = Fraction of permissible still water vertical bending moment for loading condition (j)

The fraction of permissible still water vertical bending moment β(j) for standard ship types are given in therulesStatic warping stresses are neglected for vessels with large deck openings like container vessels since thestill water torsion moment can either be positive or negative and in average zero thereby givingzero staticwarping stress in average Zero stress from horizontal static bending moment is also assumed in even keelcondition

Sec

tion

4


DNV GL AS

5 Stress due to bending of PSM

51 GeneralStresses in PSM are the results of bending of single skin or double hull structures between transversebulkheads and due to lateral pressure or loads see Figure 2 The bending stress should be determined byFE analysis or 3(2)-dimensional grillage analysis Prescriptive (analytical) methods cover a limited amount ofstructural configurationsThe stresses in PSM should be calculated for external and internal loads The loads to be used should bedetermined at the mid-position between the transverse bulkheadsIn the prescriptive assessment the bending stress of PSM should be included where relevant when relativedeflection stress is includedPrescriptive estimate of the stress of PSM from double hull bending based on analytical method is given inAppD

6 Local stiffener bending stress

61 Wave induced stiffener bending stressThe hot spot stress in Nmm2 due to local dynamic pressure in maximum and minimum load component i1and i2 (load case i) for loading condition (j) σLDik(j) is obtained from the following formula

Sec

tion

4


DNV GL AS

where

PWik(j) = Dynamic wave pressure at the mid span in kNm2 for load case i in loading condition (j) formaximum and minimum load component k (ie snap shot 1 or 2)

Pld ik(j) = Dynamic liquid tank pressure at the mid span in kNm2 for load case i in loading condition (j) formaximum and minimum load component k Pressure from different tanks acting on both sides of thestiffener should be simultaneously considered if relevant in the loading condition

Pbd ik(j) = Dynamic dry bulk cargo pressure at the mid span in kNm2 for load case i in loading condition (j) formaximum and minimum load component k

ηW ηld ηbd = Pressure normal coefficients taken as

= 1 when the considered pressure is applied on the stiffener side

= -1 otherwise

fNL = Correction factor for the non-linearity of the wave pressure as taken above

hw = Water head equivalent to the pressure at waterline in m

xe = Distance in m to the hot spot from the closest end of the span lbdg as defined in Figure 3

ℓbdg = Effective bending span of stiffener in m

Zeff-n50 = Net section modulus in cm3 of the considered stiffener calculated considering an effective flange beffof attached plating Zeff-50 is taken at the stiffener flange However for detail 32 in AppA Table 1 theZeff-50 should be calculated at a height hHS above the plate flange as shown in AppA Figure 1 Theaxial stress from the hull girder shall then also be estimated for this hot spot point rather than at thestiffener flange

beff = Effective flange in mm of attached plating specified at the ends of the span and in way of endbrackets and supports taken as above

c = ca or cf according to [77] when hot spot is located aft or forward respectively of a bulkhead andwhen relative deflection is included otherwise c = 10

Sec

tion

4


DNV GL AS

Figure 3 Definition of effective span and xe for hot spot

Sec

tion

4


DNV GL AS

62 Still water stiffener bending stressThe hot spot stress due to local static pressure in Nmm2 for loading condition (j) σLS(j) is obtained fromthe following formula

where

PS (j) = Static external pressure in kNm2 in loading condition (j)

Pls (j) = Static liquid tank pressure in kNm2 in loading condition (j)Pressure acting on both sides should besimultaneously considered if relevant in the loading condition

Pbs (j) = Static dry bulk cargo pressure in kNm2 in loading condition (j)

ηS ηlsηbs = Pressure normal coefficients taken as

η = 1 when the considered pressure is applied on the stiffener side

η = -1 otherwise

7 Local relative deflection stress

71 GeneralFor longitudinal stiffener end connections located at transverse bulkheads (or restrained frames) theadditional stress due to the relative deflection should be considered

72 Relative deflection definitionThe relative deflection is defined as the lateral displacement of the longitudinal measured at the frameforward (Fwd) or afterward (Aft) relative to the displacement at the transverse bulkhead (or restrainedframe)

73 Sign conventionThe sign of the relative deflection is positive when the stress at the hot spot is in tension The hot spot isassociated with the stress in the flange of the longitudinal at the transverse bulkhead (or restrained frame)

74 Estimate of relative deflectionThe relative deflection or stress from relative deflection can be estimated by prescriptive formulations or fromFEM analysis eg

mdash From partial ship or cargo hold FE model where the relative deflection is derived and the stress isestimated by the formulation in [75]

mdash Prescriptive method of relative deflection stress as described in AppD when applicable

A FE model is generally more applicable and should be used when available

Sec

tion

4


DNV GL AS

75 Prescriptive stress due to relative deflection derived by FE analysisThe relative deflection stress can be calculated based on relative deflection derived from FE analysis In theestimate of the relative deflection from FE analysis the local rotation at the position of the hot spot needs tobe accounted forThe stress due to relative deflection in Nmm2 for load case i1 and i2 of loading condition (j) for bothlocations lsquoarsquo and lsquofrsquo should be calculated directly using the following expression

where

a f = Suffix which denotes the location as indicated in Figure 4

Aft Fwd = Suffix which denotes the direction afterward (Aft) or forward (Fwd) from the transverse bulkhead asshown in Figure 4

Kb = Stress concentration factor due to bending for the location lsquoarsquo or lsquofrsquo which may correspond to points lsquoArsquoor lsquoBrsquo as defined in AppA

σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)

= Additional stress at location lsquoarsquo and lsquofrsquo in Nmm2 due to the relative displacement between thetransverse bulkhead including swash bulkhead or floors in way of stool and the forward (Fwd) andafterward (Aft) transverse web or floor respectively for load case i1 and i2 of loading condition (j)taken as above

IFwd-n50 IAft-n50 = Net moment of inertia in cm4 of forward (Fwd) and afterward (Aft) longitudinal

Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50

= Net section modulus of forward (Fwd) and afterward (Aft) stiffener in cm3

ℓFwd ℓAft = Span in m of forward (Fwd) and afterward (Aft) longitudinal as shown in Figure 4

xeFwd xeAft = Distance in m as shown in Figure 3 to the hot spot in location lsquoarsquo or lsquofrsquo from the closest end of lFwd

and lAft respectively

δFwdik(j)

δAftik(j)

= Relative displacement in the direction perpendicular to the attached plate in mm between thetransverse bulkhead (including swash bulkhead or floor in way of stools) and the forward (Fwd) orafterward (Aft) transverse web (or floor) as shown in Figure 4

Figure 4 Definition of the relative displacement of a longitudinal

76 Stress due to relative displacement in still waterThe additional hot spot stress in Nmm2 in still water due to the relative displacement is obtained accordingto procedures in [75] or [77] replacing the dynamic local stress σdD and dynamic pressure with static localstress σdS and static pressure

77 Consideration of reduced local stiffener bending stressWhen the length ℓAft and ℓFwd (eg at the transverse bulkhead) differ with reference to Figure 4 the localstiffener bending stress in [6] needs to be corrected at the location where the additional stress from relativedisplacement is calculated However where a stringer plate is fitted at one side of the considered web framethe local stiffener bending stress should not be corrected for the other side of the considered web frame Thecorrected stiffener bending stress due to local pressure can be calculated using correction factors ca aft andcf forward of the transverse bulkhead as follows

Sec

tion

4


DNV GL AS

8 Stress concentration factors

81 Longitudinal stiffener end connectionsThe stress concentration factors Ka and Kb are given in AppA Table 1 for end connection of stiffenerssubjected to axial and lateral loads respectivelyThe values for soft toe are valid provided the toe geometry complies with AppA [22] The stressconcentration factor Kb given for lateral loads should be used also for stress due to relative displacementsFor end connections without top stiffener eg detail 31 and 32 a recommended design is given by thedesign standard A and B in Sec8 [22] and Sec8 [21] respectively

82 Other connection types and overlapped connectionsWhen connection types other than those given in AppA are proposed the fatigue strength for the proposedconnection type may be assessed by a FE hot spot analysis as described in Sec6 to obtain directly the hotspot stress Alternatively the stress concentration factor can be calculated using FE analysis according toAppE [6]Overlapped connection types for longitudinal stiffeners ie attachments welded to the web of thelongitudinals (eg struts) are regarded as unconventional design If it is used a hot spot model is neededto establish the stress concentration factors alternatively an additional stress concentration factor of 115should be added to the K-factors in AppA

Sec

tion

5


DNV GL AS

SECTION 5 DIRECT FATIGUE STRENGTH ASSESSMENT

1 General

11 DefinitionDirect fatigue strength assessment is based on directly calculated wave loads for selected loading conditionsand for a specific wave environment The stress may be estimated by beam theory or finite element analysisReference is made to the DNVGL-CG-0130 Wave load analysis for conditions to be used in the hydrodynamicanalysis and to DNVGL-CG-0127 Finite element analysis for modelling and load application of global FEmodels

12 Stress transfer functionsThe linear hydrodynamic analysis in the frequency domain is referred to as spectral analysis The aim ofspectral fatigue calculations should establish complex stress transfer functions through direct wave loadanalysis combined with stress response analyses The stress transfer functions express the relation betweenthe wave heading and frequency and the stress response at a specific location in the structure The stresstransfer functions may be determined either by

mdash component stochastic analysis see [3]mdash full stochastic analysis see [4]

13 AssumptionsSpectral fatigue calculations imply that the simultaneous occurrence of the different load effects arepreserved through the calculations and the uncertainties are reduced compared to prescriptive calculationsThe fatigue strength assessment includes the following assumptions

mdash wave climate represented by scatter diagrammdash wave heading distribution relative to the shipmdash a speed profile which may depend on the sea state and headingmdash Probability distribution of amplitudes (eg RayleighRice) applied to the stresses within each short term

condition (sea state)mdash cycle counting consistently with the probability distribution for the short term stress responsemdash linear cumulative summation of damage contributions from each sea state in the wave scatter diagram

when the fatigue damage is calculated

14 Hydrodynamic theoryThe spectral method assumes linear load effects and responses The hydrodynamic loads should becalculated using 3D potential theory which gives a better estimate of flow field towards the ends of the shipends A benefit is also that the axial stress from the axial force component can be includedNon-linear effects due to large amplitude motions and large waves can be neglected in the fatigue analysissince the stress ranges at lower load levels dominate the fatigue damage In order to determine the rolldamping or intermittent wet and dry surfaces in the splash zone the linearization should be performed at aload level which dominates the fatigue damage

Hydrodynamic analysis is carried out to obtain transfer functions H(ωβ) of the load components whichwill be used in the component stochastic For full stochastic analysis the internal and external pressures andaccelerations (inertia) are directly transferred to the FE model to establish directly the stress transfer functionfor each node or element H(ωβ) In the transfer function ω refers to the encounter frequency and β refersto the relative heading angle between the ship and the incident waves The procedures are explained more

Sec

tion

5


DNV GL AS

in detail in DNVGL-CG-0130 Wave load analysis which covers the hydrodynamic analysis procedures and thewave environment conditions

2 Fatigue strength assessment

21 Fatigue damage calculationsWhen the long term stress range distribution is defined through a short term Rayleigh distribution withineach short term sea state the fatigue damage for bi-linear S-N curves is given in AppC [25]It should be noted that the stress in the direct calculations should account for different corrections as givenin Sec3 [25] If the stress is directly estimated from the FE then the Km factor in Sec6 [16] needs to beaccounted for in the case of welded detailsAlternative fatigue strength representations can be useful for axially loaded details or where axial loadsdominates eg

mdash Permissible peak stress range (for single loading condition)mdash Permissible stress concentration factormdash Required FAT class

as explained in Sec3 [4] and Sec3 [5]

3 Component stochastic analysisComponent stochastic calculations can be used for fatigue assessment of all details exposed to well definedload components in particular longitudinal end connections The procedure is based on combining loadtransfer functions from wave load analysis with stress factors ie stress per unit load A flow diagram of thecalculation procedure is given in Figure 1

Sec

tion

5


DNV GL AS

Figure 1 Flow diagram for component stochastic fatigue calculations

31 Load componentsThe load transfer functions to be considered can include

mdash vertical hull girder bending momentmdash horizontal hull girder bending momentmdash hull girder axial forcemdash torsional momentsmdash external (panel) pressuresmdash internal pressure due to dynamic gravity components from vessel motions and due to inertia forces from

accelerations

Load transfer functions for internal cargo and ballast fluid pressures due to accelerations in x- y- and z-direction are derived from the vessel motions

where xs ys and zs is the distance from the centre of free liquid surface to the load point in x- y- and z-direction defined by the coordinate of the free surface centre minus the coordinate of the load point

Sec

tion

5


DNV GL AS

The acceleration transfer functions are to be determined in the tank centre of gravity and include the gravitycomponent due to pitch and roll motions The gravity components are expressed as

32 From load to stress transfer functionsFor each load transfer function the corresponding stress transfer function is determined as

where

Ak = Stressload ratio for load component k

Hk(ω|b) = Load transfer function for load component k

The combined stress response is determined by a linear complex summation of the n number of stresstransfer functions

33 Stress factors per unit loadThe following stress component factors may be relevant to determine the combined stress in stiffeners andplating

A1 = Axial stress per unit vertical hull girder bending moment

A2 = Axial stress per unit horizontal hull girder bending moment

A3 = Axial stress per unit global axial force

A4 = Warping stress per unit bi-moment (from torsion) for vessels with open deck structure

A5 = Bending stress per unit local external pressure

A6 = Bending stress per unit local internal pressure (to be combined with accelerations in x- y- and z-direction)

A7 = Axial stress due to double hull bending per unit external pressure

A8 = Axial stress due to double hull bending per unit internal pressure (to be combined with accelerations in x- y- andz-direction)

A9 = Bending stress due to relative deflection per unit external pressure

A10 = Bending stress due to relative deflection per unit internal pressure (to be combined with accelerations in x- y-and z-direction)

The stress factors Ak may be either positive or negative depending on the position in the structure type ofloading and sign convention of sectional loads used in the wave load programme

Sec

tion

5


DNV GL AS

The warping stress factor A4 is related to the bi-moment which is a result of the torsional distributionTo establish the warping stress transfer function the response from the torsional distribution needs to beassessed for each period and wave heading to establish the transfer function of the bi-momentDepending on the detail to be investigated the stress per load ratio is either calculated directly byfinite element analyses or derived from the prescriptive formulas given in Sec4 combined with stressconcentration factors as given in AppA

34 Splash zone correctionIn the partly dry surface region of the ship hull (ie splash zone) as illustrated in Figure 2 the transferfunction for external pressures should be corrected to account for intermittent wet and dry surfaces Thedynamic pressure in the splash zone region PWSZ in kNm2 is estimated as

where

hW = Water head equivalent to the pressure at waterline in m

PWWL = Wave pressure at the water line in kNm2

PW = Wave pressure in kNm2 Above the water line PW is PWWL Below the water line PW is taken as thelinear hydrodynamic pressure at location z

The dynamic pressure at the mean waterline at 10-4 probability of exceedance should be used to calculatePWWL and thereby the splash zone extent Since panel pressure refers to the midpoint of the panel anextrapolation using the values for the two panels closest to the waterline has to be carried out to determinethe dynamic pressure at the waterlineAbove the waterline the pressure should be stretched using the pressure transfer function for the pressureat the waterline combined with the rp factor Below the water line but still within the splash zone the panelpressure closest to the load point should be used and this panel pressure needs to be corrected to the loadpoint by interpolation before multiplied with the rp factor Below the splash zone the panel pressure closestto the load point should be used but corrected to the load point by interpolation

Sec

tion

5


DNV GL AS

Figure 2 Reduced pressure range in the splash zone

35 Double hull and relative deflection stressStresses due to deformation of the main girder system such as double hull bending and relative deflectionare a result of the pressure distribution over the frame and girder system Consequently the correspondingstress component factors should be calculated using reference panels representative for the pressuredistribution over the hull section rather than the local pressure at the stiffener considered In general a panelat B4 from the centre line (CL) may be used as reference panel for bottom structures and a panel at 23 ofthe draught TLC may be used for side structuresRelative deflections (or double hull stresses) may be calculated by applying the directly calculated pressuredistribution based on 10-4 probability level of exceedance and including the splash zone correction tothe cargo hold FE analysis This gives a relative deflection (or double hull stress) distribution The relativedeflection (or double hull stress) distribution per unit pressure at a specific position can then be establishedThereby the relative deflection (or double hull stress) transfer function for all positions can be defined basedon the transfer function of the pressure in a single specific position To achieve increased accuracy thepressure distribution can be divided into several pressure distributions with smaller extents which can addedas load components due to linear theory

4 Full stochastic analysis

41 IntroductionIn full stochastic analysis hydrodynamic loads are directly transferred from the wave load analysis programto FE models Hydrodynamic loads include panel pressures internal tank pressures and inertia forces due torigid body accelerations Full stochastic analysis can be applied to any kind of structureThe analysis can be based on a global FE model of the vessel which can be combined with local FE modelsas sub-models As an alternative to the global model a cargo hold or partial ship FE model can be usedtransferring sectional loads calculated by the wave load program to the forward and aft end of the cargo holdor partial ship modelAll load effects are preserved through the calculations and hence the method is suitable for fatiguecalculations of details with complex stress pattern and loading Typical examples are panel knuckles bracketterminations of the main girder system larger openings and hatch corners

Sec

tion

5


DNV GL AS

A flow diagram of the calculation procedure is shown in Figure 3

Figure 3 Flow diagram for full stochastic fatigue calculations

By direct load transfer the stress response transfer functions are implicitly described by the FE analysisresults All wave headings from 0 to 360ordm with an increment of maximum 30ordm should be included For eachwave heading a minimum of 20 wave frequencies should be included to properly describe the shape of thetransfer functionsA prerequisite for correct load transfer from the hydrodynamic program is sufficient compatibility betweenthe hydrodynamic and the global FE model

mdash if different masses are used the mass distributions needs to be similarmdash if different panel model is used the buoyancy distribution needs to be similar

Similar mass properties and buoyancy can be ensured by using the structural FE model as mass model andpanel model in the hydrodynamic analysis

Sec

tion

5


DNV GL AS

Having performed the load transfer the final load equilibrium should be checked by comparing transferfunctions and longitudinal distribution of sectional forces eg bending moment and shear forces for differentwave headings Significant unbalanced forces will disturb the global response

42 Assessment of local detailsLocal FE models can be used as sub models to the global model and the displacements from the globalmodel are automatically transferred to the local model as boundary displacements In addition the localinternal and external pressure loads and inertia loads are transferred from the wave load analysisFrom the local FE models transfer functions of hot spot stress (or local stress at free plate edges) aredetermined see Sec6For details dominated by axial loading the nominal stress approach can be an efficient alternative to local FEmodels

Sec

tion

6


DNV GL AS

SECTION 6 FINITE ELEMENT ANALYSIS

1 Introduction

11 ThicknessThe thickness to be used in FE analysis is required bv the rules for the specific ship type and can be eithertn50 or tgr with the corresponding scantling approach factor fc as given in Sec3 [22] The net or grossthickness are thereby referred to as the thickness t in relation to FE analysis

12 ApplicationThis section describes the finite element (FE) method to be used for fatigue assessment The methodsare based on the hot spot stress approach as well as the local stress for free plate edge The main aim ofapplying a FE model in the fatigue analysis should obtain a more accurate assessment of the stress responseThe methods described in this Section is regarded applicable for many details using shell elements but forcertain details customized methods are regarded necessary Alternative methods are also given in AppEThree types of hot spots denoted lsquoarsquo lsquobrsquo and lsquocrsquo are described in Table 1 These are defined according to theirlocation on the plate and their orientation to the weld toe as illustrated in Figure 1 For shell elements lsquoarsquo andlsquocrsquo are for practical purposes the same while for solid elements lsquoarsquo and lsquocrsquo differs

Table 1 Types of hot spots

Type Description

a Hot spot at the weld toe on plate surface at an ending attachment

b Hot spot at the weld toe around the plate edge of an ending attachment

c Hot spot at the weld toe of an attached plate (weld toes on both the plate and attachment surface)

Figure 1 Types of hot spots

The method for calculation of hot spot stress at weld toe for any welded details is given in [41] except forweb stiffened cruciform joints The method for calculation of local stress of free plate edges is given in [42]

Sec

tion

6


DNV GL AS

The method for calculation of hot spot stress at web stiffened cruciform joints such as hopper knuckleconnection transverse bulkhead lower stool to inner bottom connection and horizontal stringer heel is givenin [5]Limitations of the hot spot stress methodology for simple connections are given in [6]

13 Stress to be used from finite element analysisThe principal stress range within 45deg of the normal to the weld toe should be used for the analysis asillustrated in Figure 2 It is assumed here that the crack is growing parallel to the weld

Figure 2 Principal stress to be used for cracks parallel to the weld

When the principal stress direction differs significantly from the normal to the weld toe it becomesconservative to use the principal stress range as indicated in Figure 2 If the angle between the principalstress direction and the normal to the weld Φ is increased further fatigue cracking may initiate in the weldand grow normal to the principal stress direction as shown in Figure 3 This means that the notch at theweld toe no longer significantly influences the fatigue capacity and a higher S-N curve or allowable hot spotstress applies for this stress direction More guidance on this effect of stress direction relative to the weldtoe as shown in Figure 2 and Figure 3 when using finite element analysis and hot spot stress S-N curves ispresented in AppE

Sec

tion

6


DNV GL AS

Figure 3 Fatigue cracking when principal stress direction is more parallel with weld toe

14 Stress field at a welded detailFigure 5 illustrated the stress increase due to the structural geometry and the notch Based on the stressfield at the surface the hot spot stress needs to be established The notch effect due to the weld is included inthe S-N curve and the hot spot stress can be derived by extrapolation of the structural stress as in [21] tothe hot spot but this Section includes the simplified method in [33]

15 Fatigue stress range based on EDW approachThe stress range to be used for each loading condition (j) is the stress range obtained from the predominantload case

where


The fatigue stress range ΔσFSi(j) in Nmm2 for free plate edge corrected for mean stress effect thicknesseffect material factor and environmental factor is taken as

For local stress at free plate edge the stress range is derived as

Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)

= Local stress in load case lsquoi1rsquo and lsquoi2rsquo of loading condition (j) in Nmm2 obtained by local stress FEanalysis

For welded details the fatigue stress range is derived as

where

ΔσFS1i(j) = Fatigue stress range due to the principal hot spot stress range ΔσHS1i(j) in Nmm2


Kp = Stress reduction factor for stress along welds given in Sec2 [24]

Km = Stress magnification factor due to misalignment (angular or due to eccentricity)

fmean1i(j)fmean2i(j)

= Correction factor for mean stress effect for the different principal stresses

fthick = Thickness effect

fmaterial = Material factor

fw = Post-weld treatment factor

fc = Scantlings correction factor

fe = Environmental factor

ΔσHS1i(j) = Principal hot spot stress ranges in Nmm2 due to the dynamic loads for load case i of loadingcondition (j) which acts within 45ordm of the perpendicular to the weld toe

ΔσHS2i(j) = Principal hot spot stress ranges in Nmm2 due to the dynamic loads for load case i of loadingcondition (j) which acts outside 45ordm of the perpendicular to the weld toe

Mean stress for base material free plate edge σmeani(j) in Nmm2 due to static and dynamic load case lsquoi1rsquoand lsquoi2rsquo of loading condition (j) is calculated asThe hot spot mean stress for welded details for loading condition j in direction 1 and 2 corresponding tonormal to weld and parallel to weld due to static and dynamic loads areestimated as

Sec

tion

6


DNV GL AS

where the index ldquo1rdquo and ldquo2rdquo after ldquoHSrdquo refers to direct normal to and parallel with the weld The hot spotstress in these cases includes both the static and the dynamic loads

16 Fatigue stress range based on direct calculationsThe stress range for each loading condition (j) is obtained

where

σFS(j) = Fatigue stress amplitude of loading condition (j) in Nmm2 obtained by hot spot FE analysis

The fatigue stress σFSi(j) in Nmm2 for welded details is given as

where

σFS1(j) = Fatigue stress amplitude due to the principal hot spot stress σHS1(j) in Nmm2


fmean1(j)fmean2(j)


σHS1(j) = Principal hot spot stress amplitude in Nmm2 due to the dynamic loads of loading condition (j) whichacts within 45ordm of the perpendicular to the weld toe

σHS2(j) = Principal hot spot stress amplitude in Nmm2 due to the dynamic loads of loading condition (j) whichacts outside 45ordm of the perpendicular to the weld toe

The fatigue stress σFS(j) in Nmm2 for free plate edge is taken as

where

σBS(j) = Fatigue stress amplitude due to the principal hot spot stress at the free plate edge in Nmm2 taken fromthe FE analysis

Mean stress for base material free plate edge or welded details σmean(j) in Nmm2 is taken from the staticloads only for loading condition (j) The mean stress should correspond to the respective stress component

Sec

tion

6


DNV GL AS

directions normal to lsquoσxrsquo or parallel to lsquoσyrsquo the weld for welded details and tangentially of the edge for freeplate edges

17 Hot spot S-N curveThe hot spot S-N curves to be used are described in Sec2 [23] For welded details S-N curve D (FAT90)should be used For the free plate edge the S-N curve depends on the surface condition

2 Finite element modelling

21 GeneralHot spot stresses and stresses at free plate edges are calculated assuming linear material behaviour andusing an idealized structural model with no fabrication-related misalignment Effects caused by fabricationimperfections as eg misalignment of structural parts must be separately accounted forModels with thin plate plane stress elements shell elements or solid elements can be used Thearrangement and type of elements have to allow for steep stress gradients and formation of plate bendingbut only the linear stress distribution in the plate thickness direction needs to be evaluated with respect tothe definition of hot spot stress This section covers shell elements

22 FE modelsThree levels of FE models are commonly used

mdash Global modelmdash Partial ship model like cargo hold model for ships with cargo holdsmdash Local models like hot spot models for welded details and local stress models for free plate edges

Description of the upper two FE models are included in the DNVGL-CG-0127 Finite element analysis Localmodels used for fatigue assessment are described in this SectionHot spot and local stress models are used for fully stochastic fatigue analyses and prescriptive fatigueanalysis Hot spot models are also being used in component stochastic analyses when the structural stressconcentration is unknown Typical considered details are

mdash Bracket toe and flange terminations of stiffenersmdash Slots and lugs in the web frames at the intersection with stiffenersmdash Bracket and flange terminations of girder systemsmdash Panel knuckles

[27] gives examples of an end bracket and a hatch corner

23 Load applicationFor prescribed fatigue calculations the loads are described by the rules Direct calculated loads are describedin Sec5 For each loading condition the following dynamic and static loads (for mean stress effect) arerelevant

mdash internal and external pressuremdash global hull girder loads for partial and local modelsmdash inertia loads of the structure and cargo in case of direct analysis

For a local model the external loads like global hull girder loads are transferred through the boundaryconditions If local loads as lateral pressure are internal loads to the local model these loads shall beincluded in the local model

Sec

tion

6


DNV GL AS

The various models have to be ldquocompatiblerdquo meaning that the larger models are to produce deformationsandor forces applicable as boundary conditions for the smaller models This stiffness in the global and localpart should be equivalent especially in the case of load transfer The extent of the local model has to bechosen such that effects due to the boundaries on the structural detail considered are sufficiently small andthat reasonable boundary conditions can be formulated

24 Mesh sizeThe evaluation of hot spot stress for lsquoarsquo type hot spot in Figure 1 should be based on shell element of meshsize t times t where t is the thickness of the plate in way of the considered hot spotThe evaluation of hot spot stress for a lsquobrsquo type hot spot in Figure 1 should be based on shell element of meshsize 10 times 10 mmThe aforementioned mesh size should be maintained within the hot spot mesh zone extending over atleast 10 elements in all directions from the fatigue hot spot position The transition of element size betweenthe coarser mesh and the hot spot mesh zone should be done gradually and an acceptable mesh qualityshould be maintained This transition mesh should be such that a uniform mesh with regular shape graduallytransitions from smaller elements to larger ones An example of the mesh transition is shown in Figure 4Mesh transitions should not be arranged close to the hot spotFor efficient read out of element stresses a mesh density in the order of t times t where t is the plate thicknessis in general preferred at the hot spot region

25 4-node or 8-node elementsFour-node shell elements with adequate bending and membrane properties are to be used inside the hotspot mesh zone The four node element should have a complete linear field of in-plane stresses and hencepure in-plane bending of the element can be exactly represented Care should be given to possible stressunderestimation especially at weld toes of type lsquobrsquo in Figure 1 Use of 4-node elements with improved in-plane bending modes is a good alternative In case of steep stress gradients 8-node thin shell elementsshould be usedThe shell elements are to represent the mid plane of the plating as indicated in Figure 7 For practicalpurposes adjoining plates of different thickness may be assumed to be median line aligned ie nostaggering in way of thickness change is required The geometry of the weld and construction misalignmentis not required to be modelledThe welds are usually not modelled except for special cases where the results are affected by high localbending e g due to an offset between plates or due to a small free plate length between adjacent weldssuch as at lug (or collar) plates Here the weld may be included by transverse plate elements havingappropriate stiffness or by introducing constrained equations for coupled node displacements A thicknessequal to two times the thickness of the plates may be used for modelling of the welds by transverse platesAll structure in close proximity to the hot spot mesh zones should be explicitly modelled with shell elementsTriangular elements close to the hot spot should be avoided where possible Use of extreme aspect ratio (egaspect ratio greater than 3) and distorted elements (eg elementrsquos corner angle less than 60deg or greaterthan 120deg) should be avoided

26 Base material free plate edgesWhere stresses are to be evaluated on a free plate edge such as hatch corners beam elements having thesame depth as the adjoining plate thickness and negligible width should be used to obtain the required localedge stress values

27 Example Hatch corners and hatch coaming end bracketIn addition to the general guidance in [21] to [26] examples of modelling a hatch corners and a hatchcoaming end bracket are illustrated in the Figure 4 and Figure 5 The hatch corner area should be meshed

Sec

tion

6


DNV GL AS

using elements with a sufficiently small size to capture the local stress on the edge The element edgedimension should not exceed 02R where R is the hatchway radius In general a minimum of 10 elements ina 90ordm arc is regarded sufficient to describe the curvature of the hatchway radius plating for a rounded cornersee Figure 5 For an elliptical or parabolic corner a minimum of 10 elements is regarded sufficient from theinboard radius end to a point on the edge located at half the longitudinal distance of the semi- major axis Atotal of 15 elements is regarded sufficient at the elliptical edge of the hatch corner see Figure 6

Figure 4 Local FE model of longitudinal hatch coaming end bracket to the deck plating with hotspot mesh zone t times t mesh

Figure 5 Example of mesh density for rounded hatch corner with 15 elements

Sec

tion

6


DNV GL AS

Figure 6 Example of mesh density for elliptical hatch corner with 20 elements

3 Derivation of hot spot stress

31 GeneralThe S-N curve to be used needs to be consistent with the meshing and the way of hot spot stressdetermination which is done by extrapolation correction andor tuning The main method explained in thissection is

mdash Derivation of stress at read out point t2 from the hot spot times a coefficient For hot spot lsquobrsquo in Figure 1read out would however be 5 mm from the hot spot

The main method is widely applicable and convenient when using the right mesh size in combination with 8-node shell elements hot spots in Figure 1 The following alternative methods can result in less conservativeresults

mdash Extrapolation from t2 and 3t2 to the hot spot see Appendix AppDmdash The main method or extrapolation but accounting for principal stress directions see Appendix AppDmdash Reduction of bending stress component for hot spot of highly localized stress peaks see [34]mdash Read out at shifted position for tuned details see [5]

In any case when shell elements are regarded non-conservative (point lsquocrsquo in Sec14) and should have beenreplaced by a solid model it can still be used provided a correction is made according to Sec6

32 Read out point at t2The stress evaluation points are located at distances t2 from the hot spot where t is the plate thicknessat the weld toe This location is denoted as the stress read out point For hot spot lsquobrsquo it would be 5 mmfrom the hot spot For modelling with shell elements without any weld the hot spot stress is taken as theprincipal stress within an angle plusmn 45o to the normal to the weld and at the read out point t2 away fromthe intersection line This 45deg is not an issue for hot spot b as the shear at the surface is zero and the axialstress can be usedThe stress components on the plate surface on which the weld is located should be evaluated

Sec

tion

6


DNV GL AS

33 Derivation of stress at read out point t2Depending on the element type one of the following stress read out methods are to be used

mdash With 4-node shell elementElement surface stress (surface where the weld is located) components at the centre points are linearlyextrapolated to the line A-A as shown in Figure 7 to determine the stress components at the stress readout point located at a distance t2 from the intersection line for type lsquoarsquo and lsquocrsquo hot spots Two principalhot spot stress ranges are determined at the stress read out point from the stress components tensordifferences (between load case lsquoi1rsquo and lsquoi2rsquo in case of EDW approach) calculated from each side (side Lside R) of line A-A The average from both sides of line A-A is taken The angle θ between the direction xof the element co-ordinate system and the principal direction pX of the principal hot spot stress range co-ordinate system has to be determined

mdash With 8-node shell elementWith a t times t element mesh using 8-node element type the element mid-side node is located on the lineA-A at a distance t2 for type lsquoarsquo and lsquocrsquo hot spots This node coincides with the stress read out pointThe element surface stress components can be used directly without extrapolation within each adjacentelement located on each side (side L side R) of the line A-A as illustrated in Figure 8 Two principalhot spot stress ranges are determined at the stress read out point from the stress components tensordifference (between load case lsquoi1rsquo and lsquoi2rsquo in case of EDW approach) calculated from each side of line A-A The average from both sides of line A-A is taken The angle θ between the direction x of the elementcoordinate system and the principal direction pX of the principal hot spot stress range coordinate systemhas to be determined

For hot spot type lsquobrsquo the derivation is simpler For fatigue assessment of type lsquobrsquo hot spots a beam elementshall be used to obtain the fatigue stress range at the read out point 5 mm from the hot spot The stressrange should be based on axial and bending stress in the beam element The beam element should havethe same depth as the connecting plate thickness while the in-plane width is negligible ie it should notcontribute to the strength and it should be compatible with the shell elements and read out the bendingstress correctlyThe above read out procedure is based on element surface stresses Generally in FE software the elementstresses are calculated at the Gaussian integration points located inside the element Depending on theelement type implemented in the FE software it may be necessary to perform several interpolations in orderto determine the actual stress at the considered stress read out point at the surface of the element mid-pointor element edge

Figure 7 Determination of stress read out points and hot spot stress for 4-node element

Sec

tion

6


DNV GL AS


34 Reduced hot spot stress for highly localized stress peaks with platebendingAt hot spots with significant plate bending in a local hot spot zone an effective hot spot stress for fatigueassessment may be derived based on the following equation

where

Δσmembrane = Membrane stress in Nmm2 at hot spot

Δσbending = Bending stress in Nmm2 at hot spot

Δσsurface = Surface stress in Nmm2 at hot spot taken from the FE analysis

The reduction factor on the bending stress can be explained by redistribution of loads to other areas duringcrack growth while the crack tip is growing into a region with reduced stress The effect is limited to areaswith a localised stress concentration which occurs for example at a hopper knuckleHowever in a case where the stress variation along the weld is small eg when the effective flange effectis insignificant the difference in fatigue life between axial loading and pure bending is also small Thereforeit should be noted that it is not correct to generally reduce the bending part of the stress to 60 This hasto be restricted to cases with a pronounced stress concentration and where the stress distribution underfatigue crack development is more similar to a displacement controlled situation than that of a load controlleddevelopment It can be applied to web-stiffened cruciform joints while the effect on other details may needto be demonstrated

Sec

tion

6


DNV GL AS

4 Procedure for analysis of standard details

41 Procedure for analysis of hot spot stress at weld toe of welded detailsFor hot spot type lsquoarsquo and lsquocrsquo the structural hot spot stress σHS is calculated from a finite element analysiswith t times t mesh density and is obtained by the following formula

where

σ = Surface principal stress in Nmm2 read out at a distance t2 away from the intersection linet = Plate thickness in mm in way of the weld toe

At structural details where the hot spot type lsquoarsquo or lsquocrsquo are classified as a web-stiffened cruciform jointthe stress read out procedure of [5] should be applied For hot spot type lsquobrsquo the stress distribution is notdependent on the plate thickness the structural hot spot stress σHS is derived from a finite elementanalysis with mesh density 10 times 10 mm and is obtained by the following formula

where

σ = Surface principal stress at edge in Nmm2 read out at an absolute distance from the weld toeof 5 mm The maximum of the surface principal stress from the two surfaces (edges) should beused

42 Procedure for analysis of local stress of base material at free plate edgeFor fatigue assessment at a free plate edge a beam element should be used to obtain the fatigue stressrange The beam element should have the same depth as the connecting plate thickness while the in-planewidth should be negligible If the free plate edge is experiencing lateral bending the beam should capturethis bending and the worst surface should be considered

5 Procedure for analysis of hot spot stress of web-stiffenedcruciform joints

51 GeneralDifferent examples of web stiffened cruciform joints are illustrated in Figure 9 Among the structural detailswhich may be relevant but not limited to the following structural details are considered as a web-stiffenedcruciform joints

mdash Welded hopper knuckle connectionsmdash Heel of horizontal stringersmdash Lower stool ndash inner bottom connection

Two kinds of hot spots relative to the web-stiffened cruciform joints should be assessed

Sec

tion

6


DNV GL AS

mdash Hot spots at the flange see [52]mdash Hot spots in way of the web see [53]

The hot spot stress for hopper knuckles of the bent type should be established by following the extrapolationprocedure described in AppE

Figure 9 Web-stiffened cruciform joints ie a 135ordm hopper knuckle a 90ordm heel in a stringer and aconnection between a deck web frame and a side web frame on a car carrier

52 Calculation of hot spot stress at the flangeFor hot spot at the flange the surface principal stress should be read out from a point shifted away from theintersection line between the considered member and the adjacent member to the position of the actual weldtoe The intersection line is taken at the mid-thickness of the cruciform joint assuming a median alignmentThe hot spot stress in Nmm2 should be obtained as

Sec

tion

6


DNV GL AS

where

σshift = Surface principal stress in Nmm2 at shifted stress read out position

The stress read out point shifted away from the intersection line is obtained as

where

t1 = Plate thickness of the plate number 1 in mm as shown in Figure 10xwt = Extended fillet weld leg length in mm as defined in Figure 10 not taken larger than t12

Figure 10 Geometrical parameters of web stiffened cruciform connections

The stress at the shifted position is derived according to the illustration in Figure 11 where α is the angle indegrees between the plates forming the web-stiffened cruciform joint

Figure 11 Procedure for calculation of hot spot stress at web stiffened cruciform connections

Sec

tion

6


DNV GL AS

The hot spot stress at the shifted position σshift is derived as

where

β = Plate angle hot spot stress correction factorσsurface(xshift) = Total surface stress in Nmm2 at xshift position (membrane stress and bending

stress)σmembrane(xshift) = Membrane stress in Nmm2 at xshift position

For α=135ordm connections the plate angle hot spot stress correction factor β is derived as

For α=120 ordm connections a correction factor is derived as


The procedure is valid for 0 le xwtt1 le 10 and the β is adjusted based on experience

Correction factors β for connections with plate angles α intermediate to those given should be derivedbased on a linear interpolation of the above values

Considering Figure 12 the surface principal stresses at the centre point of the two first elements(independent of 4- or 8-node elements) on left and right side of the line A-A are averaged and taken as thesurface principal stresses in way of the web position (line A-A) The surface principal stresses are linearlyinterpolated along the line A-A in order to determine hot spot principal stresses at the stress read out pointlocated at the xshift position as shown in Figure 12

In case of the EDW approach the two principal hot spot stress ranges are determined based on load case lsquoi1rsquoand lsquoi2rsquo

Sec

tion

6


DNV GL AS

Figure 12 Determination of stress read out points for web-stiffened cruciform connections

53 Calculation of hot spot stress in way of the webHot spots located in way of the web as indicated in Figure 13 and AppE Figure 6 are to be checked with thehot spot stress defined from the maximum principal surface stress at the intersection offset by the distancexshift from the vertical and horizontal element intersection lines as illustrated in Figure 13 The intersectionline is taken at the mid thickness of the cruciform joint assuming a median alignment The hot spot stress inNmm2 should be obtained as

where

σshift = Maximum principal surface stress in Nmm2 at the intersection offset by the distance xshift

The stress read out point at the intersection offset with reference to Figure 13 is obtained as

where

t3 = Plate thickness of the web in mm as shown in Figure 13xwt = Extended fillet weld leg length in mm taken as

Sec

tion

6


DNV GL AS

where

ℓleg1 ℓleg2 = Leg length in mm of the vertical and horizontal weld lines as shown in Figure 13

Figure 13 Hot spots in way of the web

6 Procedure for analysis of hot spot stress of simple cruciformjoints

61 Limitation to hot spot stress approachThe definition of the stress field through the plate thickness in Figure 6-4 implies that the described hot spotstress methodology is not recommended for simple cruciform joints simple T-joints or simple butt jointsAnalysing such connections with shell elements will result in a hot spot stress equal to the nominal stressThe hot spot stress approach is therefore not directly applicable when the stress flow in direction I as shownin Figure 14 is considered For stresses in the direction normal to the weld at hot spot location lsquocrsquo (directionI) there is no stress flow into the transverse plating as it is represented only by one plane in the shell modelHowever it attracts stresses for in-plane direction (direction II) at hot spot location lsquoarsquoIn situations where a bracket (or web) is fitted behind the transverse plate as shown in Figure 1 acting withstiffness in the direction normal to the transverse plate stresses flow also into the transverse plate and thehot spot methodology is considered applicable

Sec

tion

6


DNV GL AS

Figure 14 Illustration of difference to attract stresses normal to and in plane of a shell elementmodel

62 Correction of hot spot stress approach for simple cruciform jointsFor direction I at position lsquocrsquo the calculated stress from FE analysis should be multiplied by a correctionfactor K For these joints the fabrication tolerances and the corresponding stress magnification factor due tomisalignment Km are most important and need to be considered in the fatigue assessmentThe hot spot stress at position lsquocrsquo should be determined by the stress read out procedure in [3] and is takenas

where

σ = Stress read out at t2 from the hot spot in Nmm2

K = To be taken from AppA Table 8

When considering the hot spot in Figure 9 for web stiffened cruciform joints and the stress read out is takenadjacent to the centre of the web plane (where the critical hot spot is located) the detail goes from beinga web stiffened cruciform joint to a simplified cruciform join as represented by App A Table 8 It is thennecessary to consider a sufficient transverse distance from the centre of the web plane to avoid the hotspot outside of the centre of the plane to become worse than at the centre of the plane since the former iscombined with the additional factors in AppA Table 8 This could be relevant when determining the grindingextent

Sec

tion

7


DNV GL AS

SECTION 7 IMPROVEMENT OF FATIGUE LIFE BY FABRICATION

1 Introduction

11 GeneralThe fatigue performance of structural details can be improved by adopting enhanced workmanship standardswhich include building alignment and weld control Building alignment below construction tolerance couldintroduce reduced stress concentration for structural details increasing the fatigue performance Theshipbuilder is responsible to comply with the construction requirements given in the rulesThe detail design standard given in Sec8 and in DNVGL CG for specific ship types may include workmanshipand welding requirementsPost-weld fatigue strength improvement methods may be considered as a supplementary means of achievingthe required fatigue life The limitations to post-weld improvement methods are given by the rulesThe post-weld improvement method by burr grinding is applied to the weld toe Thus it is intended toincrease the fatigue life of the weld toe If the risk of failure is shifted from the weld toe to the root byapplying post-weld treatment there may be no significant improvement in the overall fatigue performanceImprovements of the weld root cannot be expected from treatment applied to weld toeThe weld notch stress concentration is a direct function of the weld flank angle and the weld toe radius Thevalidity of the S-N curves in Sec2 [2] is based on a weld flank angle with a maximum mean value of 50ordmBurr grinding will affect the weld toe radius while weld profiling will affect the weld flank angle Burr grindingis described in this section and weld profiling and other methods are given in the AppFExperience indicates that it may be a good design practice to exclude post-weld improvement at the designstage The designer is advised to improve the details locally by other means or to reduce the stress rangethrough design and keep the possibility of fatigue life improvement as a reserve to allow for possible increasein fatigue loading during the design and fabrication process

2 Grinding

21 Weld toe burr grindingThe weld may be machined using a burr grinding tool to produce a favourable shape to reduce stressconcentrations and remove defects at the weld toe see Figure 1 In order to eliminate defects such asintrusions undercuts and cold laps the material in way of the weld toe shall be removed To be efficientgrinding should extend below the plate surface in order to remove toe defectsThe total depth of the burr grinding should not exceed neither 2 mm nor 7 of the local gross thicknessof the plate Any initial undercut not complying with this requirement should be repaired by an approvedmethod

Sec

tion

7


DNV GL AS

Figure 1 Details of ground weld toe geometry

To avoid introducing a detrimental notch effect by grinding a sufficiently large burr diameter shall be usedThe diameter should be between 10 and 25 mm for application to welded joints with plate thickness from10 to 50 mm The resulting root radius of the groove shall be no less than 025 tas_built For thicker platesthan 50mm the diameter should be agreed on a case by case basis The weld throat thickness and leg lengthafter burr grinding must comply with the rule requirements or any increased weld sizes as indicated on theapproved drawingsThe treatment should produce a smooth concave profile at the weld toe with the depth of the depressionpenetrating into the plate surface to at least 05 mm below the bottom of any visible undercutTreatment of inter-bead toes should be carried out for large multi-pass welds as shown in Figure 2

Sec

tion

7


DNV GL AS

Figure 2 Extent of weld toe burr grinding to remove inter-bead toes on weld face ℓleg=ℓreg Weldleg length W Width of groove d Depth of grinding to be between 05 and 1 mm

The inspection procedure should include a check of the weld toe radius the depth of burr grinding andconfirmation that the weld toe undercut has been removed completely Grinding also improves the reliabilityof inspection after fabrication and during service lifeIf grinding is needed to achieve a specified fatigue life the hot spot stress is rather high Due to grinding alarger fraction of the fatigue life is spent during the initiation of fatigue cracks and the crack grows fasterafter initiation This may imply requirement for shorter inspection intervals during service life in order todetect the cracks before they affect the integrity of the structure

22 Flush grinding of butt weldsButt welds may be machined flush to achieve a better S-N curve with reference to AppA If a weld hasbeen machined flush it has to be documented that weld overfill has been removed by grinding and thatthe surface has been proven free from defects The weld overfill may be removed by a coarse grinder toolsuch as a coarse grit flapper disk grit size 40-60 The final surface should be achieved by fine grit grindingbelow that of weld toe defects The surface should show a smooth or polished finish with no visible scoremarks The roughness should correspond to Ra = 32 μm or better (It should be remembered that if thearea is planned to be coated a roughness around Ra = 32 μm is often recommended) The surface shouldbe checked by magnetic particle inspection It is assumed that grinding is performed until all indications ofdefects are removed Then possible presence of internal defects in the weld may be a limitation for use of abetter S-N curve and it is important to perform a reliable non-destructive examination and use acceptancecriteria that are in correspondence with the S-N classification that is used

23 Fatigue life improvement factor for burr grindingThe full effect of burr grinding is shown in Table 1 The fatigue life improvement for other methods of post-weld treatment is given in the AppF The post weld treatment factor fw can be estimated based on theimprovement factor fT as described in [24]

Sec

tion

7


DNV GL AS

Table 1 Improvement factor on fatigue life1 2)

Improvement method Minimum specified yield strength Fatigue life improvement factor fT

Less than 350 Nmm2 001ReHBurr grinding

Higher than 350 Nmm2 35

1) The improvement is only valid for high cycle fatigue where the slope m of the S-N curve is increased This isrepresented by the improvement factor on the fatigue life Alternatively the fatigue life can be assessed by using S-N curves representing the improved state when available

2) The effect of burr grinding shall not be combined with the effect of improved S-N curve through flush grinding

Sec

tion

8


DNV GL AS

SECTION 8 DETAIL DESIGN STANDARD

1 IntroductionDetail design standards are provided to ensure improved fatigue performance of structural details Thedesign standard provides fatigue resistant detail design at an early stage in the structural design processbased on the following considerations

mdash Application of fatigue design principlesmdash Construction tolerances and other practical considerationsmdash In-service experience of fatigue performance

The detail design standard may provide requirements in order to prevent the following types of fatigue cracksinitiating from the

mdash Weld toe and propagating into the base materialmdash Weld root and propagating into the plate section under the weldmdash Weld root and propagating through the weld throatmdash Surface irregularity or notch at the free plate edge and propagating into the base material

The detail design standard may cover the following

mdash Highlighting potential critical areas within the ship structuremdash Identification of the fatigue hot spot locations for each of the critical structural detailsmdash Provision of a set of alternative improved configurationsmdash Requirements on geometrical configurations scantlings welding and construction tolerancesmdash Post fabrication method for improving fatigue life

Alternative detail design configurations may be accepted if satisfactory fatigue performance or equivalenceto details in the design standard is demonstrated For details covered by a given design standard the Societymay accept that fatigue assessment by FE hot spot or local stress analysis can be omitted

2 Detail design standard for longitudinal end connections withouttop stiffener

21 Design standard A - slots with and without lug plateThis detail refers to detail 32 in AppA Table 1 and represents a longitudinal stiffener without a top stiffeneror collar plate The critical hot spots may be located in the cut outDesign standard lsquoArsquo as shown in Table 1 or equivalent is recommended for locations in way of

mdash Side shell below 11Tscmdash Bottom

The DNVGL CG for the specific ship types may also recommend design standard lsquoArsquo for locations in way of

mdash Inner hull longitudinal bulkhead below 11Tscmdash Topside tank sloping plating below 11Tscmdash Hoppermdash Inner bottom

For designs that are different from those shown in Table 1 satisfactory fatigue performance may bedemonstrated by eg using comparative FE analysis according to [23]

Sec

tion

8


DNV GL AS

Table 1 Design standard A ndash slotlug

Cut outs for longitudinals in transverse webs where web stiffeners are omitted or not connected to the longitudinal flange

Design standard A

1 2

3 4

Note 1Soft toes marked lsquorsquo are to be dimensioned to suit the weld leg length such that smooth transition from the weldto the curved part can be achieved Maximum 15 mm or thickness of transverse webcollar plateslug plates whichever isgreater

Note 2Configurations 1 and 4 indicate acceptable lapped lug plate connections

Critical location Locations around cut-out with high stress concentration and locations in way of weldterminations

Sec

tion

8


DNV GL AS


Design standard A

1 2

Detail design standard Improved slot shape to avoid high stress concentrations in transverse webs due to shearloads and local pressure loads transmitted via welded joints

Building tolerances Ensure alignment of all connecting members and accurate dimensional control of cut-outsaccording to IACS Recommendation No 47

Welding requirements A wraparound weld free of undercut or notches around the transverse web connection tolongitudinal stiffener web

22 Design standard B - collar plateThis detail refers to detail 31 in AppA Table 1 and represents a longitudinal stiffener without a top stiffenerThe critical hot spots may be located in the cut outDesign standard lsquoBrsquo as shown in Figure 1 or equivalent is recommended for locations in way of

mdash Side shell below 11 Tscmdash Bottom

The CG for the specific ship types may also recommend design standard lsquoBrsquo for locations in way of

mdash Inner hull longitudinal bulkhead below 11 Tscmdash Hoppermdash Topside tank sloping plating below 11 Tscmdash Inner bottom

For designs that are different from those shown in Figure 1 satisfactory fatigue performance may bedemonstrated by eg using comparative FE analysis according to [23]

Figure 1 Complete collar fitted before the stiffener is pulled through the web frame

Sec

tion

8


DNV GL AS

23 Equivalent design of longitudinal end connections without top stiffenerIf the designs for longitudinal end connections in [21] and [22] are not used the alternative design maybe verified to have equivalent fatigue strength to the design standard lsquoArsquo or lsquoBrsquo or may be verified to havesatisfactory fatigue performance The alternative design may be verified according to the procedure describedin the followingThe procedure is provided to verify the alternative design to have equivalent fatigue strength with respect toany position in the transverse cross section ie double bottom and double side The hot spot stress as wellas the stress at the free plate edge of the alternative design and that of the design lsquoArsquo should be comparedto each other The critical locations depend on the detail design and could be selected in agreement with theSociety The stress should be derived according to Sec6The FE models for verification of equivalence should have an transverse extent of 3 stiffeners ie 4 stiffenerspacings and the longitudinal extent should be one half frame spacing in both forward and aft direction Atypical model is shown in Figure 2 When extending this model in the longitudinal direction as in [6] themodel can also be used to study longitudinal hot spots No cut-outs for access openings are to be includedin the models Connection between the lug or the web-frame to the longitudinal stiffener web connectionsof the lug to the web-frame and free plate edges on lugs and cut-outs in web-frame are to be modelledwith elements of plate thickness size (t times t) The mesh with plate thickness size should extend at least fiveelements in all directions Outside this area the mesh size may gradually be increased in accordance with therequirements in Sec6 The eccentricity of the lapped lug plates should be included in the model Transverseweb and lug plates are to be connected by eccentricity elements (transverse plate elements) The heightof eccentricity elements are to be the distance between mid-layers of transverse web and lug plates Thethickness of the eccentricity elements should be equal to two times the thickness of web-frame plate twEccentricity elements representing fillet welds are shown in Figure 3

Figure 2 Finite element model for verification of equivalent design

Sec

tion

8


DNV GL AS

Figure 3 Modelling of eccentric lug plate by shell elements

Three load cases are to be applied to the models of the design standard and alternative designs

mdash External pressure of unit value fixed boundary conditions at top and bottom of modelmdash Shear stress by prescribed unit displacement at the model top and fixed boundary conditions at the model

bottommdash Axial load by prescribed unit displacement at the model top and fixed boundary conditions at the model

bottom

The forward and aft part of the model should have symmetry condition describing the behaviour in a doublehull structure Load application and boundary conditions are provided in Figure 4

Figure 4 Load application and boundary conditions ndash FE model for verification of alternativedesign

The alternative design may also be verified to have satisfactory fatigue performance using sub-modellingtechnique where a local model of the alternative design located at the actual position of the stiffener-frameconnection is analysed The alternative design is considered acceptable if the methodology and fatigueacceptance criterion of Sec6 is followed The alternative design is considered acceptable only for theparticular position where it is analysed

Sec

tion

8


DNV GL AS

3 Detail design standard for different ship typesThe different ship types serve different functions and the function of the vessel may define which structuraldetails that are more susceptible to fatigue damage Therefore the detail design standard for specific shiptypes are given in the DNVGL CG for the specific ship types

Sec

tion

9


DNV GL AS

SECTION 9 REFERENCES

1 References1 Det Norske Veritas Rules for Classification of Ships Part 3 Chapter 1 Hull Structural Design Ships with Length

100 Meters and above Hoslashvik January 2005

2 Hovem L Loads and Load Combinations for Fatigue Calculations - Background for the Wave Load Section forthe DNVC Classification Note Fatigue Assessment of Ships DNVC Report No 93-0314 Hoslashvik 1993

3 Cramer EH Loslashseth R and Bitner-Gregersen E Fatigue in Side Shell Longitudinals due to External WavePressure Proceedings OMAE conference Glasgow June 1993

4 Bergan P G Lotsberg I Advances in Fatigue Assessment of FPSOs OMAE-FPSO04-0012 Int Conf Houston2004

5 Berge S Kihl D Lotsberg I Maherault S Mikkola T P J Nielsen L P Paetzold H Shin C ndashH Sun H ndashH and Tomita Y Special Task Committee VI2 Fatigue Strength Assessment 15th ISSC San Diego 2003

6 British Maritime Technology BMT (Primary Contributors Hogben H Da Cunha LF and Olliver HN) GlobalWave Statistics Unwin Brothers Limited London 1986

7 Doerk O Fricke W Weissenborn C (2003) ldquoComparison of Different Calculation Methods for StructuralStresses at Welded Jointsrdquo Int J of Fatigue 25 pp 359-369

8 Fricke W (2001) ldquoRecommended Hot Spot Analysis Procedure for Structural Details of FPSOs and Ships Basedon Round-Robin FE Analysesrdquo Proc 11th ISOPE Stavanger Also Int J of Offshore and Polar Engineering Vol12 No 1 March 2002

9 Fricke W Doerk O and Gruenitz L (2004) ldquoFatigue Strength Investigation and Assessment of Fillet-Weldsaround Toes of Stiffeners and Brackets OMAE-FPSO04-0010 Int Conf Houston

10 Hobbacher A (1996) ldquoFatigue Design of Welded Joints and Componentsrdquo IIW XIII-1539-96 XV-845-96

11 Holtsmark G ldquoThe Bending Response of Laterally Loaded Panels with Unsymmetrical Stiffenersrdquo DNVC ReportNo 93-0152 Hoslashvik 1993

12 Holtsmark G Eimhjellen R and Dalsjoslash P ldquoThe Elastic Bending Response of Panel Stiffeners of UnsymmetricalCross-section subjected to Uniform Lateral Pressure Loadsrdquo DNV Report No 2004-1150 September 2004

13 Kim W S and Lotsberg I ldquoFatigue Test Data for Welded Connections in Ship Shaped Structuresrdquo OMAE-FPSO04-0018 Int Conf Houston 2004 Also Journal of Offshore and Arctic Engineering Vol 127 Issue 4November 2005 pp 359-365

14 Kuo J-F Lacey P B Zettlemoyer N and MacMillan A (2001) ldquoFatigue Methodology Specification for New-Built FPSOrdquo OMAE Paper no 3016 Rio de Janeiro

15 Lotsberg I Cramer E Holtsmark G Loslashseth R Olaisen K and Valsgaringrd S Fatigue Assessment of FloatingProduction Vessels BOSSrsquo97 July 1997

16 Lotsberg I Nygaringrd M and Thomsen T Fatigue of Ship Shaped Production and Storage Units OTC paper no8775 Houston May 1998

17 Lotsberg I and Rove H ldquoStress Concentration Factors for Butt Welds in Stiffened PlatesrdquoOMAE ASME 2000

18 Lotsberg I Overview of the FPSO Fatigue Capacity JIP OMAE Rio deJaneiro June 2001

19 Lotsberg I Design Recommendations from the FPSO Fatigue Capacity JIP PRADS Changhai 2001

20 Lotsberg I ldquoFatigue Capacity of Fillet Welded Connections subjected to Axial and Shear Loadingrdquo IIW Documentno XIII-2000-03 (XV-1146-03)

21 Lotsberg I ldquoFatigue Design of Welded Pipe Penetrations in Plated Structuresrdquo Marine Structures Vol 171 pp29-51 2004

Sec

tion

9


DNV GL AS

22 Lotsberg I ldquoRecommended Methodology for Analysis of Structural Stress for Fatigue Assessment of PlatedStructuresrdquo OMAE-FPSO04-0013 Int Conf Houston 2004

23 Lotsberg I and Sigurdsson G ldquoHot Spot S-N Curve for Fatigue Analysis of Plated Structuresrdquo OMAE-FPSO04-0014 Int Conf Houston 2004 Also Journal of Offshore and Arctic Engineering Vol 128 November2006 pp 330-336

24 Lotsberg I and Landet E ldquoFatigue Capacity of Side Longitudinals in Floating Structuresrdquo OMAE-FPSO04-0015Int Conf Houston 2004

25 Lotsberg I ldquoAssessment of Fatigue Capacity in the New Bulk Carrier and Tanker Rulesrdquo Marine Structures Vol19 Issue 1 January 2006 pp 83-96

26 Lotsberg I Rundhaug T A Thorkildsen H and Boslashe Aring (2005) ldquoFatigue Design of Web Stiffened CruciformConnectionsrdquo PRADS 2007 October 2007 Houston

27 Na J H Lee I H Sim W S and Shin H S (2003) rdquoFull Stochastic Fatigue Analysis for Kizomba lsquoArsquo FPSO-Hull Interface Designrdquo Proceedings 22nd Int conf on Offshore Mechanics and Arctic Engineering CancunMexico

28 Polezhaeva H and Chung H (2001) ldquoEffect of Misalignment on the Stress Concentration of a Welded HopperKnuckle for a Typical FPSOrdquo OMAE Rio de Janeiro

29 Sigurdsson S Landet E and Lotsberg I ldquoInspection Planning of a Critical Block Weld in an FPSOrdquo OMAE-FPSO04-0032 Int Conf Houston 2004

30 Storsul R Landet E and Lotsberg I ldquoConvergence Analysis for Welded Details in Ship Shaped StructuresrdquoOMAE-FPSO04-0016 Int Conf Houston 2004

31 Storsul R Landet E and Lotsberg I ldquoCalculated and Measured Stress at Welded Connections between SideLongitudinals and Transverse Frames in Ship Shaped Structuresrdquo OMAE-FPSO04-0017 Int Conf Houston 2004

32 Urm H S Yoo I S Heo J H Kim S C and Lotsberg I Low Cycle Fatigue Strength Assessment for ShipStructures PRADS 2004

33 Witherby amp Co Ltd 1997 Guidance Manual for Tanker Structures

34 IIW Recommendations on Post Weld Improvement of Steel and Aluminium Structures Document XIII-2200-07

35 Fricke W and Weissenborn C ldquoSimplified stress concentration factors for fatigue-prone details in shipstructuresrdquo (in German) Schiffbauforschung 43 (2004) pp 39-50

App

endi

x A


DNV GL AS

APPENDIX A STRESS CONCENTRATION FACTORS AND FAT CLASSES

1 General

11 Definition of stress concentration factorsThe fatigue life of a detail is governed by the hot spot stress range (or local stress for free plate edge) Thehot spot stress σHS is obtained by multiplying the nominal stress σn by a stress concentration factor KThe K factors in this document are thus defined as

The relation between the hot spot stress range ΔσHS to be used together with the S-N curve and thenominal stress range Δσn is

The S-N curves in Sec2 [2] are given for welded specimens where the effect of the notch is includedAll other stress risers have to be considered when evaluating the hot spot stress This can be done bymultiplication of K factors arising from different causes The K factor to be used for calculation of the hot spotstress is derived as

where

Kg = Stress concentration factor due to the gross geometry of the detail considered

Km = Stress concentration factor due to misalignment (eccentric and angular)

Kme = Stress concentration factor due to misalignment from eccentricity (used for plate butt weld connections andcruciform joints)

Kmα = Stress concentration factor due to misalignment from angular mismatch (used for plate butt weldconnections only)

Kn = Stress concentration factor for unsymmetrical stiffeners on laterally loaded panels applicable when thenominal stress is derived from simple beam analyses

For prescriptive analysis based on beam theory all the K factors may be relevant For hot spot analysis basedon FE models only Km is regarded as additional A Kg may however be relevant if the use of shell elementsis not valid see Sec6 [62] eg for simplified cruciform joints

12 Determination of stress concentration factorsThe K factors may be determined based on FE hot spot analyses as described in Sec6 Alternatively Kfactors may be obtained from tabulated factors for typical details in ships These tabulated factors can befound in [2]

App

endi

x A


DNV GL AS

13 Definition of FAT classThe FAT class refers to a detail categorization of S-N curves where the FAT number refers to the stressrange at 2middot106 number of load cycles For structural details where there is no difference between the stressconcentration from axial and bending loads or the loading is dominated by axial loads FAT class curves canbe used equivalently as K factors The FAT number given in Nmm2 is illustrated in Sec2 Figure 1

14 BasisThe K factors and FAT classes presented in the following sections cover typical details in ships Local stressconcentration in way of welds depend on the level of workmanship The default values given in the followingtables are based workmanship tolerances according to normal shipbuilding practise If greater tolerances areused the K factors or FAT classes should be determined based on the actual tolerances see also AppF

2 Longitudinal end connections

21 Tabulated K factorsKg factors for longitudinal end connections are given in Table 1 For detail 1 to 31 the hot spot is located onthe stiffener flange or top of the stiffener web in case of flat bars while for detail 32 the hot spot is locatedat the stiffener web These details are used in the prescriptive fatigue analysis in Sec4 The factors areapplicable to stiffeners subject to axial and lateral loadsWhere the longitudinal stiffener is a flat bar and there is a web stiffenerbracket welded to the flat bar thestress concentration factor listed in Table 1 (excluding detail 31 and 32) should be multiplied by a factor of112 when there is less than 8 mm clearance between the edge of the flat bar stiffener and the attachmenton either side of the attachment This also applies to unsymmetrical profiles where there is less than 8 mmclearance between the edge of the stiffener flange and the attachment eg bulb or angle profiles where theclearance of 8 mm cannot be achievedWhere the longitudinal stiffener is a rolled angle bar and there is a web stiffener (detail 1 or 2) welded tothe flange with a clearance equal to or greater than 8 mm to the edge of the stiffener flange the stressconcentration factor Kb listed in Table 1 may be multiplied by a factor of 09Designs with overlapped connectionattachments should be avoided in areas where fatigue might be criticalIf overlap is used an additional factor of 115 should be used

Table 1 Kg factors for stiffener supports

Point A Point BNo

GeometryConnection type

Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d lt150150 for150ltdlt250

160 for dgt250

114 for dle 150124

for150ltdle250134 for d gt 250

127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114

1) The attachment length d in mm is defined as the length of the welded attachment on the longitudinal stiffenerflange without deduction of the scallop

2) ID 31 and 32 refer to details where web stiffeners are omitted or not connected to the longitudinal stiffener flangeFor detail 31 however the collar is assumed welded to the stiffener flange Reduction of the Kn factor for detail 32for L and bulb profiles may be accepted based on verification by FE hot spot analysis

3) For connection type ID 32 with no collar andor web plate welded to the flange the stress concentration factorsprovided in this table are to be used irrespective of the slot configuration The fatigue assessment point lsquoArsquo is locatedat the connection between the stiffener web and the transverse web frame or lug plate The stress concentrationfactor represents the hot spot related to the longitudinal stress

For detail 32 in Table 1 the hot spot stress is not located at the flange but at the stiffener web The bendingstress can therefore be reduced compared to the bending stress in the flange by considering the hot spot tobe located a height hHS above the plate flange corresponding to the upper connection between the stiffenerweb and the frame or the lug as indicated in Figure 1 The parameter hHS should be estimated both for hotspot A and B separately If no value is available the default value of hHS can be taken as the height of thestiffener which would be conservative

App

endi

x A


DNV GL AS

Figure 1 Illustration of detail 32 with hot spot A and B in the stiffener web The arrows indicatethe stress direction A is taken at a distance hHS from the plate flange

To establish alternative K factors for actual geometries of stiffener supports the procedure in App E [6]should be followed

22 Soft toe of web stiffener and backing bracketDetail designs for longitudinal end connections with soft toes and backing brackets are given in Figure 2 Softtoe geometry of end connection of web stiffener and backing bracket is defined as

where

θ = Angle of the toe in degrees as shown in Figure 2

htoe = Height of the toe in mm as shown in Figure 2

tbkt-gr = Gross thickness of the bracket in mm

App

endi

x A


DNV GL AS

Figure 2 Detail design for soft toes and backing brackets

App

endi

x A


DNV GL AS

23 Unsymmetrical stiffener flangeThe stress concentration factor Kn for unsymmetrical flange of built-up and rolled angle stiffeners underlateral load calculated at the webrsquos mid-thickness position as shown in Figure 3 and with parameters inFigure 4 should be taken as

where

bg-n50 = Eccentricity of the stiffener equal to the distance from the flange edge to the centre line of the webin mm

bf-n50 = Net breadth of flange in mm

tf-n50 = Net flange thickness in mm

hstf-n50 = Net stiffener height including face plate in mm

tw-n50 = Net web thickness in mm

App

endi

x A


DNV GL AS

hw-n50 = Net height of stiffener web in mm

tp-n50 = Net thickness of attached plating in mm

Zn50 = Net section modulus in cm3 of stiffener with an attached plating breadth equal to the stiffenerspacing

fA = Reduction

= 05 for unsymmetrical profiles without weld at flange ie hot spot located at stiffener web (detailNo 32 in Table 1)

07 for rolled angle profiles in way of end connection of web stiffener (details No 1 - 2 in Table 1)with offset equal or greater than 8 mm

= 10 in other cases

Figure 3 Bending stress in stiffener with symmetrical and unsymmetrical flange relative tonominal stress σn

Figure 4 Stiffener - net scantling

App

endi

x A


DNV GL AS

For bulb profiles Kn factor should be calculated using the equivalent built-up profile as shown in Figure 5 Theflange of the equivalent built-up profile should have the same properties as the bulb flange ie same crosssectional area and moment of inertia about the vertical axis and neutral axis position For HP bulb profilesexamples of the equivalent built-up profile dimensions are listed in Table 2

Figure 5 Bulb profile and equivalent built-up profile

App

endi

x A


DNV GL AS

Table 2 HP equivalent built-up profile dimensions (the last row represents an approximation forHP bulb profiles for sizes not listed)

HP-bulb Equivalent built-up flange in gross thickness

Height (mm) Gross webthickness tw-gr(mm) bf(mm) tf-gr(mm) bg(mm)

200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2

hstf - tw-gr +01345hstf-21216

01317hstf- 35726

(tw-gr + 0006hstf- 03139)2

3 Butt weldsK factors and FAT classes for butt welds are given in Table 3 Default values are provided based onworkmanship tolerances according to normal shipbuilding practise If greater tolerances are used the Kfactors or FAT classes should be determined based on the actual tolerances and the equations given in Table3

Table 3 K factors and FAT classes for butt welds

No Geometry connection Description of joint Included misalignment K factor FAT Nmm2

1

Transverse butt weldwelded from both sidesground flush to plate100 NDE

Km=115 from eg axialmisalignment of 005middott(may be limited to 2 mmprovided that properquality control is carriedout1))

080 112

2

Transverse butt weldwelded from both sidesmade in shop in flatposition max weldreinforcement 1 mm +01middot weld width smoothtransitions NDE


100 90

App

endi

x A


DNV GL AS


3

Transverse butt weldwelded from both sidesand with max weldreinforcement 1 mm +015middot weld width NDE

Quality level B2)Km=13 from eg axialmisalignment of 01middott(may be limited to 3 mmprovided that properquality control is carriedout1))

113 80

Quality level C2)


125 72

4

Transverse butt weldsingle side weldingwelded on non-fusibletemporary backing rootcracking


113 80

Quality level C2)


125 72

5

Laser (t le8mm) andlaser hybrid (t le12mm)butt weld

Km=13 from eg axialmisalignment of 01middott

113 80

App

endi

x A


DNV GL AS


6

Transverse butt weldwith permanent backingstrip or three plateconnection single sidewelding

These details arenot recommended inareas prone to fatiguedue to sensitivity ofworkmanship andfabrication


127 71

Quality level C2)


142 63

7

Transverse butt weldwithout backing barsingle side welding

If root is checked byappropriate NDE FATclass (K factors) may beimproved by a factor 2

For tubular sections FATclass (K factors) maybe improved by a factor1125


248 36

Quality level C2)


278 32

8

Partial penetration buttweld

The stress should berelated to the weldthroat sectional areaand the weld overfillshould not be taken intoaccount


248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9

Transverse buttwelds of plates withangular misalignmentunsupported

The formulation assumeszero mean stressand in tension it willbecome conservativebut non-conservative incompression The α isrelated to the deflectionFor fixed ends theequation becomes equalto detail no 10

whereλ = 6 for pinnedendsλ = 3 for fixed ends

α = angular misalignmentin radians

s = plate width in mm

t = plate thickness in mm

e = deflection in mm

10

Transverse butt weldof plates or pipes withlarge radius withaxial misalignmentunsupported

where

e = eccentricity in mm


11

Transverse butt weld ofplates or pipes with largeradius with differentthickness and axialmisalignment slopedunsupported

Where

e = max misalignment

et = frac12 (t2-t1) eccentricitydue to change in thickness

t2 = thickness of thickerplate

t1 = thickness of thinnerplate

For detailno 190129middotKm3)

Fordetailsno 234amp 5

90104middotKm3)

Fordetailsno 69092middotKm3)

Fordetailsno 7amp 89047middotKm3)

For detailno 1129Km3)

Fordetailsno 234amp 5104Km3)

Fordetailsno 692Km 3)

Fordetailsno 7 amp8 47Km3)

App

endi

x A


DNV GL AS


12

Transverse butt weldof plates with differentthickness and axialmisalignment slopedsupported

Within the transversedistance 2middott2 fromthe web Outside thisdistance detail no 11should be used

Kme = 10Kmα = 10

1) In case of limited axial misalignment FAT X and K can be calculated by the relations for detail no 10 However theincreased FAT class divided by fthick is not to exceed the above tabulated FAT class and the reduced K factor times fthick isnot to be less than the above tabulated K factor respectively2) Quality level refers to the EN ISO 5817 standard3) Km should be calculated by Km=Kmα+Kme-1

4 Flange connectionsK factors and FAT classes for flange connections are given in Table 4

Table 4 K factors and FAT classes for flange connections

No Geometry Description of joint K factor FAT Nmm2

1

Flange connection with softeningtoe

147 61

App

endi

x A


DNV GL AS


2

Attachment with smoothtransition (sniped end or radius)welded on the beam flange bulbor plate

For t2 gt= 05t1 the FAT class orK factor may be improved by afactor of 1125 but not for bulb

When welding close to theedges of plates or profiles(distance less than 10mm) theFAT class and K factor shouldbe downgraded by a factor of1125

127

143

71

63

3

Longitudinal attachment weldedon edge of plate or beam flangewith smooth transition (snipedend or radius)

For t2 le 07middott1 the FAT classand K factor may be improvedby a factor of 1125

18

20

50

45

App

endi

x A


DNV GL AS


4

Longitudinal attachment weldedon edge of plate or beam flangewith smooth transition radius

Smooth transition radius formedby grinding the full penetrationweld area in order to achieve anotch-free transition area Finalgrinding should be performedparallel to the stress direction

10

127

18

90

71

50

5

Crossing of flanges welded fromboth sides and where the radiusR is ground

where

t = thickness of flange

147 61

6

Crossing flanges with fullpenetration butt weld Kme =130

180 50

App

endi

x A


DNV GL AS


7

Crossing of flanges withtransition radius R crack at freeplate edge

To be used together with S-Ncurve B to C2 dependent onsurface condition The referencestress is 19middotΔσn

The butt weld needs to beconsidered separately see no 8or 9

B-C2 150-100

8

Crossing flanged with transitionradius R full penetration buttweld welded from both sides atflat position

Welded reinforcement le 1mm+ 01middotweld width smoothtransitions NDE weld endsground Kme=115

Cutting edges in the qualityaccording to detail no 2 or no 3in Sec2 Table 2

10 90

9

Crossing flanged with transitionradius R full penetration buttweld welded from both sides

No misalignment 100 NDEweld ends ground butt weldground flush to surface

Cutting edges broken orrounded according to detail no2 in Sec2 Table 2

09 100

App

endi

x A


DNV GL AS


10

Crossing flanges with fullpenetration butt weld weldedfrom both sides Kme=130

Cutting edges in the qualityaccording to detail no 2 or 3 inSec2 Table 2

Connection length

Nominal stress

143 63

11

Crossing flanges with fullpenetration butt weld weldedfrom both sides NDE weld endsground butt weld ground flushto surface Kme=130

Cutting edges in the qualityaccording to detail no 2 and 3in Sec2 Table 2

Connection length and nominalstress as in no 10

113 80

12

Transverse butt welds betweenplates of different widths NDE

slope 15

slope 13

slope 12

In case of thickness changeadditional misalignment factorshould be considered from [3]

113

127

143

80

71

63

App

endi

x A


DNV GL AS


13

Overlap connection If this detailis being used in an area withhigh axial stress and the Kfactor of 4 is resulting in fatiguelife below the target the actualK factor should be estimatedbased on FE analysis

40 23

5 Welded attachments

51 Attachments welded to a plate or stiffenerK factors and FAT classes for attachments welded to a plate or stiffener are given in Table 5

Table 5 K factors and FAT classes for attachments welded to a plate


1

Transverse attachment on unsupportedplate Edge distance if any to beminimum 10mm

t le 25 mm (Kmα=12)

t gt 25 mm and

tp gt 25mm (Kmα=12)

tp gt 50mm and tp gt 2t (Kmα=112)

tp = Base plate thickness

For transverse attachment supportedbelow see detail no 4 in Table 8

113

127

119

80

71

76

2

Longitudinal attachment welded on beamflange bulb or supported plate

For t2 le 05t1 FAT class (K factors) maybe improved by a factor of 1125 but notbetter than FAT 80 (K=113) (not valid forbulb profiles)

113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3

Fillet welded non-load-carrying overlapjoint at a axially loaded component

-Flat bar

-Bulb profile

-Angle profile

For l gt 150mm the FAT class (K factor)has to be downgraded by a factor of1125 while for l le 50mm the FAT class(K factor) may be improved by a factorof 1125 If the component is subject tobending the FAT class (K factor) has to bedowngraded by a factor of 1125

160

160

180

56

56

50

4

Fillet welded overlap joint with smoothtransition (sniped end with φ le 20deg or radius) welded to axially loadedcomponent

-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5

Longitudinal attachment welded on edge ofplate or beam flange

For t2 le 07t1 FAT class (K factor) maybe improved by a factor of 1125 but notbetter than FAT 56 (K=16) If the plateor beam flange is subjected to in-planebending the FAT class (K factor) has to bedowngraded by a factor of 1125

160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6

Stud welding on a plate or bulb profile Foran adequate workmanship on bulb profile acentric connection is required

Non-load carrying

Load carrying

113

20

80

45

52 Termination of stiffenersK factors and FAT classes for termination of stiffeners are given in Table 6

Table 6 K factors and FAT classes for termination of stiffeners


1

Local attachments weldedto unsupported plates egbuckling stiffener

θ = angle in degrees of slopingtermination

90K

App

endi

x A


DNV GL AS


2

Sniped stiffener flange

The nominal stress at the hotspot should be considered

L profile should be consideredwith care due to twisting of thestiffener from lateral loadingcausing local bending stress inthe web close to clamped ends(as illustrated based on a topview of a L profile)

90K

53 Doubling platesK factors and FAT classes for doubling plates are given in Table 7

Table 7 Doubling Plates


1

End of doubling plate on beam welded ends(based on stress range in flange at weld toe)The values are referring to and doubling platethickness in the range of 08tplt tD le15tpd le 50

50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d

The following features improves the FAT class (Kfactor) by a factor of 1125

-reinforced ends or

-weld toe angle le 30deg

For tD le 08tp the FAT class (and K factor) shouldbe improved by a factor of 1125For tD gt 15tp the FAT class (and K factor) shouldbe downgraded by a factor of 1125

For all improvements the FAT class is limited by80 corresponding to a K factor of 113

120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2

Doubling plates welded to unsupported platesd le 50

50 lt d le 100

100 lt d le 150

150 lt d le 300

For doublers with d gt 300 a more detailedanalysis should be performed based on the actualgeometry

127

132

143

161

71

68

63

56

If the welds of the doubling plates are placed closer to the free plate edges than 10 mm the FAT classes and K factorsshould be downgraded by a factor 115

6 Simple cruciform jointsK factors and FAT classes for cruciform joints are given in Table 8

Table 8 K factors and FAT classes for cruciform joints


1

Cruciform or tee-joint K-butt welds with fullpenetration or defined incomplete root penetrationThe effective weld thickness may be assumedas the thickness of the abutting plate t1 minusf where f is the incomplete root penetration of02t1 with a maximum of 3 mm which should bebalanced by equally sized double fillet welds oneach side e0 le 03t1 Kmα = 104) Kme =1453) θ= 45deg1)

Cruciform joint

Tee-joint

127

113

71

80

2

Cruciform or tee-joint with transverse fillet weldstoe failure (for root failure in particular for throatthickness alt07t1 or alt07t2 see detail no 3) eaccording to detail no 1 e0 le 03t1 Kmα = 104)Kme =1453) θ = 45deg2)

Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3

Cruciform or tee-joint with transverse fillet orpartial penetration welds root failure based onstress range in weld throat (for toe failure seedetail no 2) e according to detail no 1 e0 le03t1 Kmα = 104) Kme =1453)

The stress in the weld throat can be estimatedbased on the nominal stress in the member withthickness t1

σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4

Cruciform joint or double sided transverseattachment with non-load carrying transverseattachment with thickness t With full penetrationor fillet weld Kmα = 1124) Kme =10 θ = 45deg1)

t le 25 mm

t gt 25 mm

105

119

86

76

5

Simple (unsupported) cruciform joint with supportshaving no rotational fixation e is the eccentricityof plates with thickness t1 and t2 t is the thicknessof the considered plate In case the vertical plate iscontinuous t3=t4

Reference todetail no 1to 3

Referenceto detailno 1 to3

App

endi

x A


DNV GL AS


1) For weld angle deviating from 45deg the K should be multiplied by c=112middot(06+06tan(θ)14) For correction of the FATclass the FAT class should be multiplied with 1c

2) For weld angle deviating from 45deg the K should be multiplied by c=1167middot(08+087tan(θ)14) For correction of theFAT class the FAT class should be multiplied with 1c

3) If the misalignment is different than the default value for detail no 1 to no 3 then K factor can be multiplied withKmemodKme and the FAT class can be multiplied with KmeKmemod

4) For angular misalignment reference is made to detail no 9 in Table 3 If both angular misalignment and misalignmentdue to eccentricity is considered then Km should be calculated by Km=Kmα+Kme-1

7 Scallops and block jointsK factors and FAT classes for scallops are given in Table 9 The factors are applicable to stiffeners subjectedto axial loads For dynamic pressure loads on the plate other design solutions may be necessary

Table 9 K factors and FAT classes for scallops

No Geometry Description of joint K factor12) FAT12) Nmm2

1

Circular scallop with fullpenetration butt weld weldedfrom both sides misalignment isnot included (in Point A)

Point A

Point B

24

127

375

71

2

Scallop with full penetration buttweld welded from both sidesmisalignment is not included (inpoint A)

Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3

Scallop with full penetration buttweld welded from both sidesmisalignment is not included (inPoint A)

Point A

Point B

117

127

77

71

4

Omega shaped scallop with fullpenetration butt weld weldedfrom both sides misalignment isnot included (in Point A)

Point A

Point B

117

127

77

71

NotesFor scallops without transverse welds the Kg at point B will be governing for the design

For Ω shaped scallops with different shape than in no 4 and assessment based on the hot spot analysis may be used todemonstrate value for point A which may be worse than for no 41) If the scallop height is more than 40 of the web height then FAT class and K factor should be downgraded by afactor of 11252) In the presence of shear stress τ in the web the FAT class and K factor should be downgraded by the factor 1-ΔτΔσbut not below FAT 36 or above K of 25

8 Lower hopper knucklesK factors and FAT classes for lower hopper knuckles with angles (between inner bottom and hopper plate)between 30deg and 75deg are given in Table 10 and serves as references although they are regarded to dependon geometry and scantlingsFor the lower hopper knuckle the nominal stress to be used with the K factors and FAT classes in Table 10should be taken as the transverse membrane stress Its stress should be read out frac12 a stiffener spacing fromthe knuckle in the inner bottom plate The stress should be taken as the averaged of the two stresses in themiddle between the adjacent two floors It is assumed that brackets (extension plate) are fitted in ballasttanks in line with inner bottom plate and that a longitudinal girder (or similar brackets) are located beneaththe lower hopper knuckleGeometrical eccentricity in the knuckle should be avoided or kept to a minimumFor a yard standard geometry a K factor related to nominal stress in a frame and girder model mayalternatively be established using the hot spot model A hot spot stress calculation of the panel knuckles mayalso be required by the rules

App

endi

x A


DNV GL AS

Table 10 K factors and FAT classes for lower hopper knuckles


1

Lower hopper knuckle without insert plate orbracket and with transverse weld at the hotspot which is caused by the high effectiveflange effect

70 13

2

Insert plate of two times the thickness normallyrequired Insert plates should be provided ininner bottom hopper tank top and web frameThe insert plates should extend approximately400 mm along inner bottom and hopper plateapproximately 800 mm in longitudinal directionand 400 mm in the depth of the web

45 20

3

Bracket inside cargo tank The bracket shouldextend approximately to the first longitudinalsand the bracket toe should have a soft nosedesign

25 36

App

endi

x A


DNV GL AS

9 Pipe connectionsK factors and FAT classes for pipe connections are given in Table 11

Table 11 K factors and FAT classes for pipe connections


1

Full penetration weld at the connection betweena hollow section (eg pillar) and a plate

For tubular section

For rectangular hollow section

For t le 8mm the FAT class and K factor shouldbe downgraded by a factor of 1125

161

180

56

50

2

Fillet weld at the connection between a hollowsection (eg pillar) and a plate

For tubular section


The stress should be related to the weldsectional area For t le 8mm the FAT class andK factor should be downgraded by a factor of1125

200

225

45

40

3

Continuous butt or fillet weld connecting a pipepenetrating through a plate

d le 50 mm

d gt 50 mm

For large diameters an assessment based onlocal stress is recommended or alternativelyfollow [10]

127

143

71

63

10 Holes with edge reinforcementsK factors for holes with reinforcement are given for the following details

mdash Holes in plates with inserted tubular are given in Figure 7 to Figure 18mdash Holes in plates with ring reinforcement are given in Figure 19 to Figure 23mdash Holes in plates with double ring reinforcement are given in Figure 24 to Figure 27

App

endi

x A


DNV GL AS

For stresses parallel with the weld the given stress concentration factors can be reduced according to Sec2[24] Table 3

At some locations of the welds there are stress in the plate transverse to the fillet weld σw and shear stressin the plate parallel with the weld τlll Then the fillet weld is designed for a combined stress obtained as

where

t = plate thickness in mma = throat thickness in mm for a double sided fillet weld

Figure 6 Kg at hole with inserted tubular Stress at outer surface of tubular parallel with weldHtr = 2

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS

Figure 8 Kg at hole with inserted tubular Stress in plate parallel with weld Htr = 2

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS

Figure 10 Kg at hole with inserted tubular Stress in plate normal to weld Htr = 2

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS

Figure 12 Kg at hole with inserted tubular Principal stress in plate Htr = 2

Table 12 Angle to principal stress Htr=2

trtp rtp=10 rtp=20 rtp=50 rtp=100

00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS

Figure 14 Kg at hole with inserted tubular Shear stress in plate Htr = 2

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS

Figure 18 Kg at hole with ring reinforcement Max stress concentration

App

endi

x A


DNV GL AS

Figure 19 Kg at hole with ring reinforcement Stress at inner edge of ring

App

endi

x A


DNV GL AS

Figure 20 Kg at hole with ring reinforcement Stress in plate parallel with weld

App

endi

x A


DNV GL AS

Figure 21 Kg at hole with ring reinforcement Shear stress in weld

App

endi

x A


DNV GL AS

Figure 22 Kg at hole with ring reinforcement Stress in plate normal to weld

App

endi

x A


DNV GL AS

Figure 23 Kg at hole with double ring reinforcement Stress at inner edge of ring

App

endi

x A


DNV GL AS

Figure 24 Kg at hole with double ring reinforcement Stress in plate parallel with weld

App

endi

x A


DNV GL AS

Figure 25 Kg at hole with double ring reinforcement Shear stress in weld

App

endi

x A


DNV GL AS

Figure 26 Kg at hole with double ring reinforcement Stress in plate normal to weld

11 Rounded rectangular elliptical and oval holesK factors for rounded rectangular holes are given in Figure 27 and K factors for elliptical and oval holes aregiven in Figure 28

App

endi

x A


DNV GL AS

Figure 27 K factors for rounded rectangular holes

App

endi

x A


DNV GL AS

Figure 28 K factors for elliptical and oval holes

In case of superimposed stresses due to longitudinal and shear loads the maximum local stress σloc in Nmm2 of rectangular openings with rounded corners can approximately be calculated as follows

where

m = Parameter according to Figure 29

Kt = Local stress factor for equivalent stress

App

endi

x A


DNV GL AS

c = Parameter according to Figure 29

ρ = Ratio of smaller length to radius of corner defined as ρ=ℓr or ρ=ar with ρ gt=3

ℓ = Length of opening in m

a = Height of opening in m

τ1 = Shear stress related to gross area of section in Nmm2

σ1 = Longitudinal stress (in direction of length ℓ of opening) related to gross area of section in Nmm2

r = Radius of rounded corner in m

App

endi

x A


DNV GL AS

Figure 29 Parameters m and c to determine the local stress factors of rectangular openingsloaded by superimposed longitudinal and shear stresses 35

App

endi

x B


DNV GL AS

APPENDIX B FATIGUE CAPACITY

1 Failure criterion inherent in the S-N curvesMost S-N data are derived by fatigue testing of small specimens in test laboratories For these specimens thetesting is performed until final fracture In comparison to real ship structures no redistribution of stressesoccurs during crack growth This means that most of the fatigue life is associated with growth of a smallcrack which grows faster as the crack size increases until fractureFor details with the same calculated damage the initiation period of a fatigue crack takes longer time for arounded notch in base material than at the notch of a weld toe or a weld root (which are normally sharper)This also means that with a higher fatigue resistance of the base material as compared with welded detailsthe crack growth will be faster in the base material when fatigue cracks are growing as the stress range inthe base material can be significantly higher than at the welds if they are designed with the same fatigueutilizationFor practical purpose one defines these failures as being crack growth through the thicknessWhen this failure criterion is transferred into a crack size in a real structure where some redistribution ofstress is more likely this means that this failure criterion corresponds to a crack size that is somewhat lessthan the plate thickness

2 Other S-N curves

21 S-N curves for base material of high strength steel above 500Nmm2The fatigue capacity of the base material is depending on the surface finish of the material and the yieldstrength For high strength steel other than cast steel with yield strength above 500 Nmm2 and a surfaceroughness equal Ra = 32μm or better the following design S-N curve can be used for fatigue assessment ofthe base material

The mean S-N curve is given by log N = 17770 - 470 log Δσ In air a fatigue limit at 2middot106 cycles at astress range equal 235 Nmm2 can be used For variable amplitude loading with one stress range largerthan this fatigue limit a constant slope S-N curve should be usedFor seawater with cathodic protection aconstant slope S-N curve should be used (The same as for air to the left of 2middot106 cycles see Figure 1) Ifrequirements to yield strength surface finish and corrosion protection are not met the S-N curves presentedin Sec2 [23] should be used The thickness exponent is n = 0 for this S-N curve

App

endi

x B


DNV GL AS

Figure 1 S-N curve for high strength steel (HS - curve)

22 Qualification of new S-N curves based on fatigue test dataOther S-N curves than given by this CG can be applied subject to the following conditionsFor qualification of new S-N data it is important that the test specimens are representative for the actualfabrication and construction This includes possibility for relevant production defects as well as fabricationtolerances The sensitivity to defects may also be assessed by fracture mechanicsFor new types of connections it is recommended to perform testing of at least 15 specimens in order toestablish a new S-N curve At least three different stress ranges should be selected in the relevant S-Nregion such that a representative slope of the S-N curve can be determined Reference is made to theDNVGL-RP-0005 Fatigue design of offshore steel structures D7 for a broader assessment of how to deriveS-N curves from few fatigue test dataReference is also made to IIW document no IIW-XIII-WG1-114-03 for statistical analysis of the fatigue testdata Normally fatigue test data are derived for number of cycles less than 107 However it should be notedthat for ship structures significant fatigue damage occurs for N gt 107 cycles Thus how to extrapolate thefatigue test data into this high cycle region is important in order to achieve a reliable assessment procedureIn addition to statistical analysis engineering judgement based on experience is relevant for derivation of theS-N data in this region Good details where the fatigue initiation contributes significantly to the fatigue lifeshow a more horizontal S-N curve than for less good details where the fatigue life consists mainly of crackgrowth Reference is also made to S-N curves with different slopes shown in Sec2 [2]The residual stresses of small scale test specimens are often small compared to residual stress in full sizestructures due to different restraints during fabrication This is an item which is of importance when planningfatigue testing and for assessment of design S-N curves Reference is made to the commentary section D7in DNVGL-RP-0005 Fatigue design of offshore steel structuresFor N gt 107 cycles there is additional uncertainty due to variable amplitude loading This is an issue if it isdesirable to establish a less conservative S-N curves than given in this CGThe detection probability of defects during production should be considered The defects with acceptabledetection probability are normally larger in real structures than for test specimens

App

endi

x B


DNV GL AS

3 Effect of fabrication tolerancesLarger fabrication tolerances are allowed in real structures than accounted for in the test specimens used toderive S-N data ref DNVGL-OS-C401 Fabrication and Testing of Offshore Structures Therefore additionalstresses resulting from normal fabrication tolerances should be included in the fatigue design Specialattention should be given to the fabrication tolerances for simple butt welds in plates and cruciform joints asthese give the most significant contributions Stress concentration factors for butt welds and cruciform jointsare given in AppA

4 Design chart for fillet and partial penetration weldsDesign should be performed such that fatigue cracking from the root is less likely than from the toe regionFatigue crack at the toe can be found by in-service inspection while a fatigue crack starting at the root cannot be discovered before the crack has grown through the weld Thus the design of the weld geometryshould be performed such that the fatigue life for cracks starting at the root is longer than the fatigue life ofthe toe Figure 3 can be used for evaluation of required weld penetration The notation used is explained byFigure 2 It is difficult to detect internal defects by NDE in filletpartial penetration welds Such connectionsshould therefore not be used in structural connections of significant importance for the structural integrity

Figure 2 Welded connection with partial penetration weld

App

endi

x B


DNV GL AS

Figure 3 Weld geometry with probability of root failure equal toe failure

App

endi

x C


DNV GL AS

APPENDIX C FATIGUE DAMAGE CALCULATIONS AND FATIGUESTRESS DESIGN TABLES

1 Weibull long term stress distributionThe long term distribution of stress ranges may be described by the two-parameter Weibull distribution

where

Q = probability of exceeding the stress range Δσ

ξ = Weibull slope parameter

q = Weibull scale parameter

The stress range distribution may also be expressed as

where

Δσo = reference (largest) stress range value at the local detail exceeded once out of no cycles

no = total number of cycles associated with the stress range level Δσo

The Weibull slope parameter is established based on long term stress or wave load analysis The Weibulldistribution is a fit to the stress range distribution at different probability levels of exceedance The Weibullslope depends on the location of the detail and the governing load components The reference level for thestress is not significant if the Weibull slope and scale parameters are well established Alternatively thefatigue strength may be assessed by considering the different encountered sea states and summarise theshort term results from all the sea states

2 Fatigue damage accumulation

21 Practical damage calculation with one slope S-N curveThe long term stress range distribution can be expressed by a stress histogram consisting of a number ofconstant amplitude stress range blocks Δσi each with a number of stress repetitions ni Based on this thefatigue criterion reads

App

endi

x C


DNV GL AS

where

D = accumulated fatigue damage

m K2 = S-N fatigue parameters

ntot = number of stress blocks

ni = number of stress cycles in stress block i

Ni = number of cycles to failure at constant stress range Δσi

η = Maximum acceptable usage factor

22 Closed form damage estimate for one-slope S-N curves and WeibulldistributionThe fatigue damage can be estimated based on the Weibull long term stress distribution and the one-slope S-N curve For one loading condition the fatigue damage is given by

where

ND = Total number of stress cycles experienced by ship during the design fatigue life taken as ND =31557times106 (f0TD) (4 log L)

f0 = Factor taking into account time in seagoing operations

TD = Design life in years

ν0 = average zero-crossing frequency of the stress for the specific detail = 1(4log(L)) based on wavebending moment

Γ (1+mξ) = Gamma function Values of the gamma function are listed in Table 1

Use of one slope S-N curves give conservative estimates of the fatigue lives when using the slope at N lt 107

cycles

App

endi

x C


DNV GL AS

Table 1 Numerical values for Γ(1+mξ)

ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306

23 Closed form damage estimate for two-slope S-N curves and WeibulldistributionThe fatigue damage can be estimated based on the Weibull long term stress distribution and the two-slope(bi-linear) S-N curve For one loading condition the fatigue damage is given by

where

Δσq = Stress range in S-N curve where the change of slope occurs

K2 m = S-N fatigue parameters for N lt 107 cycles

K3 m+Δm = S-N fatigue parameters for N gt 107 cycles

γ( ) = Incomplete Gamma function to be found in standard tables

Г( ) = Complementary Incomplete Gamma function to be found in standard tables

24 Closed form damage estimate for one slope S-N curve and short termRayleigh distributionThe fatigue damage can be estimated based on the one-slope S-N curve short term Rayleigh distributionwithin each short term sea state all sea states and all loading conditions in the following way

App

endi

x C


DNV GL AS

where

rinj = the relative number of stress cycles in short term condition i n for loading condition j

αj = Probability of loading condition j

nLC = Total number of loading conditions

moinj = zero spectral moment of stress response process in short term condition i n and loading condition j

25 Closed form damage estimate for two-slope S-N curve and short termRayleigh distributionThe fatigue damage can be estimated based on the two-slope S-N curve short term Rayleigh distributionwithin each short term sea state all sea states and one loading conditions in the following way

3 Maximum allowable stress rangeThe maximum allowable stress range Δσ0 results for different Weibull shape parameters ξ three S-N curves(DFAT90 CFAT125 and B2FAT140 in air) three probabilities of exceedance (10-2 10-4 and 10-8) and threedesign lives (given by 05middot108 07middot108 and 10middot108 cycles) are given in Table 2 to Table 4The maximum allowable stress range can be assumed to include K factors The maximum allowable nominalstress range can then be obtained as

Alternatively if you assume that K=1 then the tables provide the maximum allowable hot spot or local stressrangeExample

mdash Weibull shape parameter ξ = 10mdash total number of stress cycles = 07middot108

mdash welded joint in-air environment

It follows from Table 2 to Table 4 that the maximum allowable stress range is

mdash 75 Nmm2 at 10-2 probability level of exceedancemdash 150 Nmm2 at 10-4 probability level of exceedance andmdash 300 Nmm2 at 10-8 probability level of exceedance

App

endi

x C


DNV GL AS

If the stress concentration factor is K=15 then the maximum nominal stress range at a probability level ofexceedance of 10-2 is 51 Nmm2 which can be used for screening

In lieu of more accurate calculations the slope parameter ξ for vertical wave bending moment may betaken as

ξ = Weibull slope parameter for wave bending moment

=

As another example the vertical bending is a crucial design load for cruise vessels For a vessel rule lengthL between 250 and 350m the Weibull slope parameter is between 092 and 084 The corresponding zeroup-crossing period is between 96 to 102sec For a life time of 25 years this gives 078108 to 082108

cycles Based on a probability level of 10-2 10108 cycles and a free plate edge with normal workmanship(CFAT class 125) Table 2 gives the rule of thumb value of 100MPa for the stress rangeIn Table 5 the maximum allowable stress range is estimated for all the S-N curves listed in Sec2 Table 1 Themaximum allowable stress range is based on a 10-2 probability level of exceedance and a Weibull slope of ξ =10

Table 2 Maximum allowable stress range (Nmm2) at a probability of exceedance 10-2 to keep thefatigue damage less than 10 for different design life cycles

Welded joint DFAT90 In-Air Δσ0

Free plate edge CFAT 125 and B2140 In-air Δσ0

Weibull Shape-parameter ξ

Design life cycles Design life cycles

05108cycles 07108cycles 10108cycles 05108cycles 07108cycles 10108cycles

060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS




Weibull shape-parameter ξ



060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS

Table 5 Maximum allowable stress range (Nmm2) at a probability of exceedance 10-2 and forWeibull shape parameter ξ = 10 to keep the fatigue damage less than 10 for different design lifecycles and S-N curves

In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45

4 Guidance for omission of fatigue analysisA detailed fatigue analysis can be omitted if the largest hot spot stress range or local stress range for actualdetails in air or cathodic protected environment is less than the fatigue limit at 107 cyclesThe use of the fatigue limit is illustrated in Figure 1 A detailed fatigue assessment can be omitted if thelargest stress cycle is below the fatigue limit However in the example in Figure 2 there is one stress cycleΔσ1 above the fatigue limit This means that a further fatigue assessment is required This also meansthat the fatigue damage from the stress cycle Δσ2 has to be included in the fatigue assessment and thesummation of fatigue damage presented in this document should be used

App

endi

x C


DNV GL AS

Figure 1 Stress cycling where further fatigue assessment can be omitted

Figure 2 Stress cycling where a detailed fatigue assessment is required

App

endi

x D


DNV GL AS

APPENDIX D PRESCRIPTIVE FATIGUE STRENGTH ASSESSMENT

1 Bending stress of PSM

11 Prescriptive double hull bending stress at transverse bulkheadWhen relative deflection (see [2]) contributes then also double hull bending stress may contribute Doublehull bending contributes when the double hull is represented by a grillage system with longitudinal girders(stringers) The following prescriptive estimates serve as an alternative to more advanced FE analysis Figure1 gives an illustration of the main dimensions of the grillage model with longitudinal stress due to double hullbending In Figure 2 the girder system with longitudinal stiffening is illustrated The stiffeners are smearedout in the prescriptive model to contribute to the flange area of the longitudinal girders (stringers)

Figure 1 Double hull configuration with illustration of double hull bending stress (Sa=S Sb = l)

App

endi

x D


DNV GL AS

Figure 2 Double hull configuration with distances between girders (t=tn50 Sa=S Sb = ℓ)

12 Prescriptive wave induced double side bending stress at the transversebulkheadThe longitudinal double side hot spot stress (or nominal stress with Ka=1) σdhDik(j) in Nmm2 is given as

where

Ka = Axial stress concentration factor from AppA for the respective detail

Kc = Coefficient dependent on aspect ratio ρ

Ke = Coefficient for loading area Ke = 10 for double side and bottom

ra = Distance from neutral axis of the double hull panel to the point considered in mm defined as positiveoutwards towards the outer shell

Kd = Coefficient for loading area Kd = TLCD for double side and Kd = 10 for double bottom

a = Longitudinal length of double hull panel (bottom or side) in mm

b = Transverse width of double hull panel (bottom or side) in mm

ia = Average moment of inertia per longitudinal double panel girder about transverse neutral axis of thegirder taken as ia= IaS in mm3

App

endi

x D


DNV GL AS

ib = Average moment of inertia per transverse double panel girder about longitudinal neutral axis of thegirder taken as ib= Ibl in mm3

Ia Ib = Longitudinal and transverse moment of inertia about the double hull neutral axis in transverse andlongitudinal axes respectively in mm4 per girder including effective flange In the computation of Ia thearea of the longitudinal stiffeners may be added to the plate as a smeared out thickness

Sl = Spacing between girders in longitudinal and transverse direction respectively in mm

PWik(j) = Dynamic pressure in kNm2 for external pressure at TLC for double side and at B4 for double bottomThe pressure should be taken at a global x-position corresponding to the bulkhead considered

Pldik(j) = Dynamic internal pressure (from the cargo tank) in kNm2 on the longitudinal bulkhead If the cargotank is not carrying fluid the internal pressure should be neglected The pressure should be taken at aglobal x-position corresponding to the longitudinal centre of the tank located on the same side of thetransverse bulkhead as the detail considered at a transverse position corresponding to the longitudinalbulkhead and at a vertical position corresponding to the vertical centre of the tank

z = Vertical distance from base line in m to the considered detail

D = Depth moulded in m

It is assumed that the grillage is clamped at the ends (due to symmetry) and simply supported at thelongitudinal edges The Kc can then be estimated as

13 Prescriptive still water double side bending stress at the transversebulkheadThe longitudinal double side hot spot stress (or nominal stress with Ka=1) σdhS(j) in Nmm2 can be givenas

where

Kd = Coefficient for loading area Kd = TLCD2 for double side and Kd = 10 for double bottom

Ke = Coefficient for loading area Ke = 12 for double side and Ke = 10 for double bottom

PS (j) = Static external pressure in kNm2 in loading condition (j) and taken at z=0 ie at the base line Thepressure should be taken at a global x-position corresponding to the bulkhead considered

Pls (j) = Static internal pressure (from the cargo tank) in kNm2 in loading condition (j) If the cargo tank isnot carrying fluid the internal pressure should be neglectedThe pressure should be taken at a global x-position corresponding to the longitudinal centre of the tank located on the same side of the transversebulkhead as the detail considered and at a vertical position corresponding to the tank top

App

endi

x D


DNV GL AS

2 Local relative displacement stress

21 Prescriptive wave induced relative deflection for double hull vesselsAs an alternative to relative deflection estimated by FE analysis the wave induced relative deflection hot spotstress amplitude σdDik(j) in Nmm2 for load case ik and loading condition (j) may be estimated as follows

where

xe = Distance in m to the hot spot from the closest end of the span ℓbdg as defined in Sec4 Figure 3

z01 = Distance from neutral axis of the stiffenerplate to stiffener flange in mm

Zn50 = Net section modulus for the stiffenerplate in mm3 referring to the stiffener between the transversebulkhead and the adjacent frame on which the hot spot is located

ℓbdg = The effective length of the stiffener in m

E = Youngrsquos Modulus in Nmm2

In50 = Net moment of inertia in mm4 for the stiffenerplate referring to the stiffener between thetransverse bulkhead and adjacent frame on which the hot spot is located

dik(j) = Relative deflection in mm for load case ik and loading condition j

The formulation is regarded applicable for transverse bulkheads without transverse stringers restrainingthe adjacent frame An illustration of the relative deflection for the adjacent frame is given in Figure 3 Therelative deflection dik(j) in mm for the side shell can be estimated as

where

dmik(j) = Maximum relative deflection in mm over the height of the web frame for load case ik and loadingcondition j



The maximum relative deflection is estimated as

App

endi

x D


DNV GL AS

where

S = Sum of half plate flange width on each side of the horizontal stringer closest to D2 in mm

ηl = Sign factor being 1 for stiffener in outer shell (ship side and bottom) and -1 for stiffener in inner sideor inner bottom

ℓ = Frame spacing in mm

Ia = Moment of inertia for the longitudinal stringer in mm4

Ib = Moment of inertia for the transverse frame in mm4

Ns = Number of cross ties

ia ib = IaS Ibℓ

PWik(j) = Dynamic external pressure in kNm2 at TLC The pressure should be taken at a global x-positioncorresponding to the bulkhead considered

Pldik(j) = Dynamic internal pressure (from the cargo tank) in kNm2 on the longitudinal bulkhead If thecargo tank is not carrying fluid the internal pressure should be neglected The pressure should betaken at a global x-position corresponding to the longitudinal centre of the tank located on the sameside of the transverse bulkhead as the detail considered at a transverse position corresponding tothe longitudinal bulkhead and a vertical position corresponding to the z=D2

App

endi

x D


DNV GL AS

Figure 3 Illustration of stress due to relative deflection between a transverse bulkhead and theadjacent frame

22 Prescriptive relative deflection for double hull vessels due to staticpressureThe static hot spot stress amplitude for relative deflection σdS(j) in Nmm2 is given as

App

endi

x D


DNV GL AS

where

PS (j) = Static external pressure in kNm2 in loading condition (j) and taken at z=0 ie at the base lineThe pressure should be taken at a global x-position corresponding to the bulkhead considered


Pls (j) = Static internal pressure (from the cargo tank) in kNm2 in loading condition (j) If the cargo tankis not carrying fluid the internal pressure should be neglected The pressure should be taken ata global x-position corresponding to the longitudinal centre of the tank located on the same sideof the transverse bulkhead as the detail considered at a transverse position corresponding to thelongitudinal bulkhead and a vertical position corresponding to the tank top

Ke = Coefficient from loading area Ke = 12 for double side and Ke=10 for double bottom

3 Local plate bending stress

31 Wave induced plate bending stressThe longitudinal plate bending stress at the plate short edge σPLD is located at the weld of the platetransverse framebulkhead intersection midway between the longitudinals The wave induced hot spot stressamplitude (or nominal stress with K=1) in load case i1 and i2 for loading condition (j) σPLDik(j) in Nmm2 isgiven by

where

K = Stress concentration factor which depends on the plate thickness in AppA If nothing else is specifiedK can be taken as 113 for fillet welds

s = Stiffener spacing in mm

tn50 = Net plate thickness in mm

The transverse plate bending stress at the plate long edge σPTD is located at the weld at the stiffener mid-length The wave induced hot spot stress amplitude (or nominal stress with K=1) in load case i1 and i2 forloading condition (j) σPTDik(j) in Nmm2 is given by

App

endi

x D


DNV GL AS

For local plate bending due to local pressures the following sign conventions apply

mdash Positive for pressure acting on the welded side of the plate (tension at hot spot)mdash Negative for pressure acting on the non-welded side of the plate (compression at hot spot)

32 Still water plate bending stressThe longitudinal and transverse static stress depends on the loading condition (j) The longitudinal static hotspot stress amplitude (or nominal stress with K=1) for loading condition (j) σPLS(j) in Nmm2 is given by

The transverse static hot spot stress (or nominal stress with K=1) for loading condition (j) σPTS(j) in Nmm2 is given by

App

endi

x E


DNV GL AS

APPENDIX E FINITE ELEMENT ANALYSIS

1 FE modelling

11 Mesh sizeFor 8-node shell elements and 4-node shell elements with additional internal degrees of freedom forimproved in-plane bending behaviour a mesh size from t2 up to 2t may be used For conventional 4-nodeelements a mesh size from t2 to t may be used Larger mesh sizes at the hot spot region may provide non-conservative results

12 Solid elementsAn alternative particularly for complex cases is offered by solid elements These need to have adisplacement function capturing steep stress gradients eg by using linear plate bending stress distributionin the plate thickness direction This is offered eg by iso-parametric 20-node elements (with mid-sidenodes at the edges) which mean that only one element in plate thickness direction is required An easyevaluation of the membrane and bending stress components is then possible if a reduced integration orderwith only two integration points in the thickness direction is chosen A finer mesh sub-division is necessaryparticularly if 8-node solid elements are selected Here at least four elements are recommended in thethickness direction Modelling of the welds is generally recommended as shown in Figure 1For modelling with solid elements the dimensions of the first two or three elements in front of the weld toeshould be chosen as follows The element length may be selected to correspond to the plate thickness Inthe transverse direction the plate thickness may be chosen again for the breadth of the plate elementsHowever the breadth should not exceed the ldquoattachment widthrdquo ie the thickness of the attached plate plustwo times the weld leg length (in case of type c in Figure 1 The thickness of the web plate behind plus twotimes the weld leg length) The length of the elements should be limited to 2middott It is recommended that alsothe fillet weld is modelled to achieve proper local stiffness and geometryIn order to capture the properties of bulb sections with respect to St Venant torsion it is recommendedto use several solid elements for modelling of the bulb section The meshing of the bulb however has tobe adopted to the meshing at the relevant hot spots which are located at the end connection in way if theframe


21 Extrapolation from t2 and 3t2Recommended stress evaluation points are located at distances t2 and 3middott2 away from the hot spot wheret is the plate thickness at the weld toe These locations are also denoted as stress read out points Theextrapolation method is regarded as the basic method but read out at t2 with stress correction is regardedmore convenient for many standard detailsFor modelling with shell elements without any weld considered the following procedures can be used

mdash A linear extrapolation of the stresses to the intersection line based on read out points at t2 and 3t2 fromthe intersection line The principal stress at the hot spot is calculated from the extrapolated componentvalues (principal stress within an angle plusmn 45o to the normal to the weld)

For modelling with solid elements with the weld included the following procedures can be used

mdash A linear extrapolation of the stresses to the weld toe based on read out points at t2 and 3t2 fromthe weld toe The principal stress at the hot spot is calculated from the extrapolated component values(principal stress within an angle plusmn 45o to the normal to the weld)

App

endi

x E


DNV GL AS

The stress components on the plate surface should be evaluated along the paths shown in Sec6 Figure 14and Figure 1 and extrapolated to the hot spot The average stress components between adjacent elementsare used for the extrapolation

Figure 1 A FE model with solid elements with stress extrapolation to the weld toe

22 Derivation of effective hot spot stress considering principal stressdirectionsThe S-N curves defined in Sec2 [2] are derived based on stress range normal to the weld As the anglebetween the principal stress direction and the normal to the weld is increased it becomes conservative touse the principal stress range An alternative method for deriving the hot spot stress is described below andcan replace the procedure described in Sec6 [32]Two alternative methods can be used for hot spot stress derivation Method A and method BMethod AThe linear extrapolation for shell and solid elements are according to [21] The notations for stresscomponents are shown in Figure 2 and Figure 3 The effective hot spot stress σEff in Nmm2 to be usedtogether with the hot spot S-N curve is derived as

where the principal stresses are calculated as

App

endi

x E


DNV GL AS

whereKp= 090 if manual fillet or butt welds are carried outKp = 080 if automatic fillet or butt welds are carried out from both sides with stop and start positionsKp = 072 if automatic fillet or butt welds are carried out from both sides without stop and start positionsc = 10The equation for effective stress accounts for the situation with fatigue cracking along a weld toe as shownin Figure 2 and fatigue cracking when the principal stress direction is more parallel with the weld toe asshown in Figure 3 This is then a refinement compared to Sec6 [14] where only the principal stress moreparallel to the weld included a Kp factor The relation above is more cumbersome as also the weld type needto be identifiedMethod BFor modelling with shell elements without any weld the hot spot stress is taken as the stress at the read outpoint t2 from the intersection line For modelling with solid elements with the weld included in the modelthe hot spot stress is taken as the stress at the read out point t2 from the weld toeThe effective hot spot stress is derived as in Method A but with the parameter c = 112

3 Derivation of stress at read out points

31 Derivation of stress at read out points t2 and 3t2The stress at the read out points is established as described in the following Alternatively the nodal stressesmay be used provided that they are derived directly from the calculated element stresses within eachelement When referring to hot spot lsquoarsquo and lsquocrsquo in Figure 1 t2 and 3t2 are used but when considering hotspot lsquobrsquo 5 and 15 mm are used as basis4-node shell elements t2 le element size le t

mdash element surface stress at the centre points are used as illustrated in Figure 2 a)mdash the stress at the element centre points are extrapolated to the line A-A as shown in Figure 2 b) to

determine the stress at read out pointsmdash if the mesh density differs from t times t the stresses at the stress read out points are determined by 2nd

order interpolation as shown in Figure 2 c)

8-node shell elements t2 le element size le t

mdash element edge surface stress points can be used as stress read out points as illustrated in Figure 3 a) andb)

mdash if the mesh density differs from t times t the stresses at the read out points are determined by 2nd orderinterpolation as shown in Figure 3 b)

8-node shell elements t le element size le 2t

mdash element surface result point stress is used as illustrated in Figure 4 a)mdash the stress at the surface result points are extrapolated to the line A-A by 2nd order interpolation as shown

in Figure 4 b)mdash the stress at the read out points are determined by 2nd order interpolation as shown in Figure 4 c)

Solid elements

mdash in case of solid elements the stress may first be extrapolated from the Gaussian points to the surfaceThen these stresses can be interpolated linearly to the surface centre or extrapolated to the edge of theelements if this is the line with the hot spot stress derivation to determine the stress read out points

App

endi

x E


DNV GL AS

a)

b)

c)

Figure 2 Determination of stress read out points and hot spot stress for 4-node shell elements

App

endi

x E


DNV GL AS

a)

b)

Figure 3 Determination of stress read out points and hot spot stress for 8-node shell elementst2le element size le t

App

endi

x E


DNV GL AS

a)

b)

c)

Figure 4 Determination of stress read out points and hot spot stress for 8-node shell elements t leelement size le 2t

App

endi

x E


DNV GL AS

4 Procedure for analysis of hot spot stress for tuned details

41 Bent hopper tank knucklesThe hot spots at the inner bottomhopper sloping plate in transverse and longitudinal directions are referredto as hot spot 1 2 and 3 on the ballast tank side in Figure 5 The hot spot stress of a bent hopper knuckleshould be taken as the surface principal stress read out from a point shifted from the intersection line Theshifted read out position is defined in Sec6 [52] The hot spot stress σHS in Nmm2 is obtained as

where

σbending(xshift) = Bending stress in Nmm2 at xshift positionσmembrane(xshift) = Membrane stress in Nmm2 at xshift position

The procedure that applies for hot spots on the ballast tank side of the inner bottomhopper plate in way of abent hopper knuckle is in principal the same as that applied on the cargo tank side of the inner bottom platefor welded knuckle in Figure 10 and Figure 12 The intersection line is taken at the mid-thickness of the jointassuming median alignment The plate angle correction factor and the reduction of bending stress as appliedfor a web-stiffened cruciform joint in Sec6 [52] are not to be applied for the bent hopper knuckle type

The stress at hot spots located in way of the transverse web and side girder ie hot spots 4 5 and 6 definedin Figure 6 of a bent hopper knuckle type should be derived as described for web-stiffened cruciform jointsin Sec6 [53]

App

endi

x E


DNV GL AS

Figure 5 Hot spots for inner bottom plate of bent lower hopper tank knuckle

App

endi

x E


DNV GL AS

Figure 6 Hot spots for attached supporting girders of bent lower hopper tank knuckle

5 Screening procedure

51 IntroductionA screening procedure can be used to assess the fatigue strength of specified structural details It is basedon screening stress at weld toes and the hot spot stress is determined by multiplying the stresses obtainedfrom local FE models by a stress magnification factor η of the considered detail

52 ProcedureThe screening procedure includes the following three phases

mdash Phase 1 Calculation of fatigue stress

mdash Stresses are calculated at the stress read out point from the local FE analysis with elements size of 50x 50 mm according to DNVGL-CG-0127 Finite element analysis for all fatigue load cases and for allloading conditions Stresses to be used are element average membrane components stress defined in[56]

mdash Hot spot surface stress components are calculated for each load case lsquoi1rsquo and lsquoi2rsquo from the stressesmultiplied by the stress magnification factor η taken as

σHS i1(j) = η σS i1(j)


mdash Hot spot principal surface stress ranges are the difference of hot spot stress components obtained foreach load case lsquoi1rsquo and lsquoi2rsquo

mdash Fatigue stress ranges for welded joints are determined from hot spot principal surface stressranges with correction factors as according to Sec6 [1] (eg mean stress effect thickness effectenvironmental factor and scantling approach factor)

whereσS i1(j) Stress calculated from the local FE analysis in load case lsquoi1rsquo of loading condition (j) defined in[54] or [55]σS i2(j) Stress calculated from the local FE analysis in load case lsquoi2rsquo of loading condition (j) defined in[54] or [55]

App

endi

x E


DNV GL AS

η Stress magnification factor given in Table 1

mdash Phase 2 Selection of S-N curveThe S-N curve D defined in Sec2 [2] should be used with the fatigue stress range of weld toe from thescreening

mdash Phase 3 Calculation of fatigue damage and fatigue life according to Sec3 [3]

Similar screening can be based on directly calculated loads in Sec5 rather than prescriptive loads from therules

Table 1 Stress magnification factor η for bulk carriers and oil tankers

Ship type Structural details category Stress magnification factor η

Toe of stringer 245Oil tanker

Bracket toe of transverse web frame 165

Lower hopper knuckle 210 for FA1)200 for EA1)

Lower stool ndash inner bottom (for knuckle angle = 90deg)

166

Bulk carrier

Lower stool ndash inner bottom (for knuckle angle gt 90deg)

145 for FA1)175 for EA1)

1) FA and EA mean full and empty cargo hold in alternate loading condition respectively

53 Screening fatigue criteriaThe total fatigue damage and the fatigue life of screened details should comply with the criteria given in therulesStructural details which do not comply with the acceptance criteria can be checked with respect to fatiguestrength using a FE hot spot analysis as described in Sec6 or alternatively the screening may be repeatedbased on upgraded scantlings

54 Stress read out procedure for bracket toeFor bracket toe the stress read out point is located 50 mm from the bracket toe as shown in Figure 7

App

endi

x E


DNV GL AS

Figure 7 Stress read out point at bracket toe

55 Stress read out procedure for knuckle detailFor the lower hopper knuckle and for the connection between transverse bulkhead lower stool and innerbottom the stress read out point is located 50 mm from the knuckle line (ie model intersection line) asshown in Figure 8

Figure 8 Stress read out point of knuckle detail

App

endi

x E


DNV GL AS

56 Read out point stressThe average of membrane stress components at the centre of the four adjacent elements to the stressread out point (or node) modelled with elements size of 50 x 50 mm should be used as read out pointstress When the element size is less than 50 x 50 mm the stress at the read out point can be derived usingelements in an equivalent area as shown in Figure 9

Figure 9 Equivalent area for element size less than 50 x 50 mm

App

endi

x E


DNV GL AS

6 Derivation of stress concentration factors for alternative endconnection designsThe structural stress concentration factors for alternative designs should be calculated by a FE hot spotmodel analysis according to Sec6 Additional procedure for derivation of geometrical stress concentrationfactors for stiffener end connections using FE hot spot model analysis are given below

a) FE model extent The FE model as shown in Figure 10 should cover at least four web frame spacingsin the longitudinal stiffener direction and the considered detail should be located at the middle frameThe same type of end connection should be modelled at all the web frames In the transverse directionthe model may be limited to one stiffener spacing (half stiffener spacing on each side) For L profilesthree stiffener spacings is regarded necessary to limit twisting or alternatively coupling where the twolongitudinal free plate edges should have the same rotation and displacement

b) Load application In general two loading cases are to be considered

mdash Axial loading by enforced displacement applied to the model ends andmdash Lateral loading by unit pressure load applied to the shell plating

c) Boundary conditions

mdash Symmetry conditions are applied along the longitudinal cut of the plate flange (T profiles and flatbars) and symmetry conditions along transverse and vertical cuts on web frames and on top of theweb stiffener

mdash For lateral pressure loading The model should be fixed in all degrees of freedom at both forward andaft ends as well as at the transverse cut at the top of the three web frames

mdash For axial loading The model should be fixed for displacement in the longitudinal direction at the aftend of the model while enforced axial displacement is applied at the forward end or vice versa

d) FE mesh density At the location of the hot spots under consideration the element size should be in theorder of the thickness of the stiffener flange In the remaining part of the model the element size shallbe in the order of s10 where s is the stiffener spacing

Figure 10 Fine mesh finite element model for derivation of geometrical stress concentrationfactors

For the two loading cases the stress concentration factors are determined as follows

App

endi

x E


DNV GL AS

σHSa = Hot spot stress in Nmm2 determined at the stiffener flange for the axial displacement

σna = Nominal axial stress in Nmm2 calculated at the stiffener flange based on an a unit axialdisplacement applied at one end of the FE model

σHSb = Hot spot stress in Nmm2 determined at the stiffener flange for the unit pressure load

σnb = Nominal bending stress in Nmm2 calculated at the stiffener flange according to Sec4 [6] in way ofthe hot spot for the unit pressure load applied to the FE model

The nominal stress should be determined consistently with the expressions used to derive the nominalstress from beam theory to ensure consistency between the stress concentration factor and the prescriptiveapproach

7 Verification of analysis methodologyThe verification of the hot spot stress methodology based on hot spot FE models may be carried out forspecific details with already derived target hot spot stress Such details with target hot spot stress are shownin the Commentary section of DNVGL-RP-0005 Fatigue design of offshore steel structures

App

endi

x F


DNV GL AS

APPENDIX F WORKMANSHIP AND LINK TO ANALYSIS PROCEDURES

1 Tolerance limitsThe fatigue life of a detail depends on the workmanship in terms of geometrical imperfections internaldefects and external defects due to the welding process and surface preparation of base material and freeplate edgesInternal and external defects are weld toe undercuts cracks overlaps porosity slag inclusions andincomplete penetration Geometrical imperfections are defined as axial misalignment (eccentricity) angulardistortion excessive weld reinforcement and otherwise poor weld shapesFollowing common workmanship practise embedded internal defects like porosity and slag inclusion are lessharmful for the fatigue strength compared to external defectsAppA gives equations for calculation of K factors and FAT classes due to fabrication tolerances formisalignment of butt welds and cruciform joints as well as effects of the local weld geometry The defaultvalues for K factors and FAT classes including misalignments according to common workmanship standard aregiven in AppA Some default values are given in Table 1 The S-N curves given in Sec2 [2] include the effectof the notch at the weld or free plate edge as well as internal and external defects according to commonworkmanship standardIf the fabrication tolerances related to equations for misalignment in AppA are exceeded by poorworkmanship or a higher workmanship standard is followed modified K factors and FAT classes can beestimated A higher workmanship standard may be difficult to achieve and follow up during production andthe Society may require additional documentation for acceptance of such a standard

Table 1 Default fabrication tolerances

TYPE OF IMPERFECTION TYPE OF IMPERFECTION EMBEDDEDIMPERFECTIONS SURFACE IMPERFECTIONS

POROSITY ISOLATED2) Maxpore diameter dMin distance toadjacent pore

Internal and externaldefects t4 max 4 mm25d 3 mm25d

POROSITY CLUSTERED1) Max porediameter dMax length of cluster Internal defects 3 mm25 mm Not applicable

SLAG INCLUSIONS1)2)3) MaxwidthMax length WeldDiscontinuities 30 mmt max 25 mm Not applicable

UNDERCUTMax depth(Smoothtransition required) WeldDiscontinuities Not applicable 06 mm

UNDERFILL1)2) Max depthMaxlength WeldDiscontinuities Not applicable

15 mmt2

EXCESSIVE WELDREINFORCEMENT2)4)Max height Geometrical Imperfections Not applicable b5 max 6 mm

OVERLAP1)2) Max Length WeldDiscontinuities Not applicable t

CRACKS WeldDiscontinuities Not accepted Not accepted

LACK OF FUSION WeldDiscontinuities Not accepted Not accepted

LINEAR MISALIGNMENT Maxeccentricity butt joints Maxeccentricity cruciform joints

Geometrical Imperfections Not applicable See AppA Table 3 andAppa Table 8

App

endi

x F


DNV GL AS


ANGULAR MISALIGNMENTMaxdistortion Geometrical Imperfections Not applicable See AppA Table 3 and

AppA Table 8

INCOMPLETE PENETRATION1)2)

Max lengthMax heightWeld

Discontinuitiest15 mm

tt10 max 15 mm

1) Defects on a line where the distance between the defects is shorter than the longest defect are to be regarded asone continuous defect

2) t Plate thickness of the thinnest plate in the weld connection3) If the distance between parallel slag inclusions measured in the transverse direction of the welding is less than 3

times the width of the largest slag inclusion the slag inclusions are regarded as one defect4) b Width of weld reinforcement5) IACS Rec No 476) ISO 5817

2 Weld profiling by machining and grindingWeld profiling is referred to as profiling by machining or grinding Profiling by welding is not consideredefficient to improve the fatigue strengthIn design calculations the thickness effect for cruciform joints T-joints and transverse attachments maybe reduced to an exponent n = 015 provided that the weld is profiled by either machining or grinding to aradius R of approximately half the plate thickness t2 with stress direction as shown in Figure 1When weld profiling is performed with geometric parameters are given in Figure 1 a reduced hot spot stress(or nominal stress at hot spot) σr in Nmm2 due to reduced notch stress can be calculated as

The local stress (hot spot stress or nominal stress at hot spot) should be separated into a membrane partand a bending part

where

σmembrane = Stress in Nmm2 at middle surface at location of the hot spot

σbending = Deduced bending stress component in Nmm2 at upper surface at location of the hot spot

σlocal = Hot spot stress (or nominal stress at hot spot) in Nmm2 at upper surface at location of hot spot

App

endi

x F


DNV GL AS

Figure 1 Weld profiling of load carrying and non-load carrying cruciform joints

3 TIG dressingThe fatigue life may be improved by TIG dressing as given in Table 2 Due to uncertainties regarding qualityassurance of the welding process this method may not be recommended for general use at the design stage

4 Hammer peeningThe fatigue life may be improved by means of hammer peening as given in Table 2 However the followinglimitations apply

mdash Hammer peening should only be used on members where failure will be without substantial consequencesmdash Overload in compression must be avoided because the residual stress set up by hammer peening will be

destroyedmdash It is recommended to grind a steering groove by means of a rotary burr of a diameter suitable for the

hammer head to be used for the peening The peening tip must be small enough to reach weld toe

Due to uncertainties regarding quality assurance of the process this method may not be recommendable forgeneral use at the design stage

5 Improvement factors for TIG dressing and Hammer peeningImprovement factor for toe grinding is given in Sec7 [2] while improvement factors for TIG dressing andhammer peening are given in Table 2 The factor assumes full effect of the post weld treatment method butthe acceptance of considering the full effect depends on agreement by the Society The post weld treatmentfactor fw can be estimated based on the improvement factor fT as described in Sec3 [24]

App

endi

x F


DNV GL AS

Table 2 Improvement factor on fatigue life1 3 4)


Less than 350 Nmm2 001ReHTIG dressing


Less than 350 Nmm2 0011ReH for constant amplitude loadingHammer peening2)



2) The improvement effect depends on the tool used and workmanship If lacking experience with respect to hammerpeening it is recommended to perform fatigue testing of relevant detail (with and without hammer peening) beforea factor on improvement is decided

3) The technique are normally not accepted for ships4) The improvement factor can not be claimed in addition to improvement factor for any form of grinding

App

endi

x G


DNV GL AS

APPENDIX G UNCERTAINTIES IN FATIGUE LIFE PREDICTIONS

1 IntroductionThere are a number of uncertainties associated with fatigue life predictions These uncertainties are relatedto both the wave loading and to the fatigue capacityFatigue damage is proportional to stress raised to the power of the inverse slope m of the S-N curve iea small change in the stress results in a much greater change in the fatigue life Special attention shouldtherefore be given both to stress raisers from workmanship and to the loadingThere is a large uncertainty associated with the determination of S-N curves The scatter in the test resultswhich form the basis for the S-N curves is related to the variation of the fabrication quality The ratiobetween calculated fatigue lives based on the mean S-N curve and the mean minus two standard deviationsS-N curve is significant as shown in Figure 1 The latter is used for designThere is also uncertainty associated with the determination of the stress concentration factor The errorintroduced in the calculated fatigue life by wrong selection of the stress concentration factor is indicated inFigure 2

Figure 1 Fatigue life influence of stress level and S-N data for welded connections

App

endi

x G


DNV GL AS

Figure 2 Fatigue life sensitivity to stress concentration factor K and Weibull shape factor ξ

2 Probability of fatigue failureA high predicted fatigue life is associated with a reduce probability of fatigue failure see Figure 3 It mayreduce the need for in-service inspection A high calculated fatigue life means that the accumulated fatiguedamage occurring during the service life is in the left part of Figure 3Reliability methods may be used to illustrate the effect of uncertainties on the probability of a fatigue failureReference is made to Figure 4 which shows accumulated probability of a fatigue failure as function ofyears in service for different assumptions of uncertainty in the input parameters The left part of Figure 4corresponding to the first 20 years service life is shown in Figure 5Figure 3 and Figure 4 shows accumulated probability of fatigue failure for uncertainty in S-N datacorresponding to a standard deviation of 020 in logN scale A normal distribution in logarithmic scale isassumed The uncertainty in the Miner summation assumed log-normal distributed with median 10 and CoV(Coefficient of Variance) equal to 03 (30) Uncertainties to the load and response are assumed normaldistributed with CoVnom equal 15-20 and hot spot stress derivation is assumed normal distributed withCoVhs equal 5-10

App

endi

x G


DNV GL AS

Figure 3 Calculated probability of fatigue failure as function of calculated damage

App

endi

x G


DNV GL AS

Figure 4 Accumulated probability of fatigue crack as function of service life for 20 years designlife

App

endi

x G


DNV GL AS

Figure 5 Accumulated probability of fatigue crack as function of service life for 20 years designlife (left part from Figure 3)

3 Size of defects inherent in S-N data and requirements to non-destructive examinationMost fatigue cracks are initiated from undercuts at weld toes There can be different reasons for undercuts Itis recommended to prepare a good welding procedure to avoid large undercuts during production welding Noundercut is included in S-N curves better than D (FAT 90) or where the hot spot curve D is associated witha K factor less than 10 Maximum allowable size of undercuts equal to 05 mm in depth may be consideredinherent in the different design S-N classes D E and F (FAT 90 to FAT 71) For stricter S-N curves (FAT below71) an undercut equal to 10 mm may be acceptedThe weld details can be subjected to larger stress ranges when better S-N curves are used Butt welds withlower K factors (higher FAT classes) are more critical with respect to internal defects than many other detailsPlanar defects like cracks and lack of fusion are not allowed It should be noted that use of S-N curves betterthan D (K factor below 1 or FAT class above 90) imply special consideration with respect to requirements toNDE and acceptance criteria Thus the requirements in ISO 5817 are not strict enough for details belongingto S-N class better than D Reference is made to NORSOK M-101 and ISO 5817 regarding requirements tovolumetric defects like porosity and slag inclusions

App

endi

x G


DNV GL AS

An example of crack growth through a plate thickness of 25 mm based on an initial maximum surface defectsare shown in Figure 6 based on different S-N curves A semi-elliptic surface defect is assumed with half axesac = 02 (a = crack depth c = crack half width) This calculation is based on a Weibull long term stressrange distribution with 108 cycles during a life time of 20 years The shape parameter is ξ = 10 and thescale parameter is determined such that the accumulated Miner damage is equal 10 during the life time ofthe structure It is observed that larger surface defects can be accepted for S-N curve F (K=127FAT71)as compared with that of D (K=10FAT90) and C (K=08FAT112) In Figure 6 the initial crack depths areestimated to 128 mm for F 023 mm for D and 037 mm for C The reason for the low value for the S-Ncurve D as compared with C is the notch effect at the weld toe which is not present for a machined groundflush surface C The situation becomes different for internal defects where only very small defects areacceptable for a C class due to high stresses inside the weld

Figure 6 Crack growth development from maximum surface defects in some different S-N classes

App

endi

x H


DNV GL AS

APPENDIX H LOW CYCLE FATIGUE

1 GeneralThe low cycle fatigue (LCF) procedure is based on frequent loading and unloading cycles The procedure isbased on the required design life TD If extended design fatigue life TDF is required the number of staticcycles and dynamic cycles inducing high cycle fatigue (HCF) should be adjusted accordinglyDetails will experience static and dynamic loads during their life time Fatigue strength of details may berequired based on HCF loads Even though HCF has been checked at the design stage some cracks havebeen reported within few years of operation For such cases LCF due to repeated yielding in compression andtension have been the causeLCF strength should be assessed for highly stressed locations under cyclic static loads as repeated yieldingcan cause cracks andor paint cracks at hot spots even though the dynamic stress is lowA fatigue life in low cycle high stress region is expressed in terms of the total strain range rather than thestress range An approach based on the pseudo-elastic hot spot stress range is adopted This approach iscompatible with the hot spot strain range approach as total strain is converted to pseudo-elastic stress rangeby using a plasticity correction factorAn S-N curve approach in the LCF region below 10000 design cycles is usedThe following locations may be vulnerable in view of LCF

mdash Web stiffener on top of inner bottom longitudinal and hopper slope longitudinals when wide frame space isemployed

mdash Web-frame hotspots at the stiffener-frame connections in areas of high girder shear stress or where webstiffener is not fitted on top of longitudinal flange

mdash Heel and toe of horizontal stringer of transverse bulkhead for frequent alternate loadingmdash Inner bottom connection to transverse bulkhead for frequent alternate loadingmdash Lower stool connection to inner bottom for a loading condition with one side tank empty and the other

tank fullmdash Any other locations under repeated high static stress ranges

The procedure is developed for the LCF assessment with the following limitations

mdash new building of steel ship structuresmdash steel materials with yield stress less than 355 Nmm2

mdash same LCF performance for base metal and welded jointsmdash the maximum principal stress direction does not change for a load condition

The procedure describes how to calculate combined fatigue damage due to LCF and HCF for free plate edgesand welded details The combined fatigue damage due to HCF and LCF should be below acceptable limits Thesimplified LCF procedure is shown in Figure 1

App

endi

x H


DNV GL AS

Figure 1 Assessment procedure for low cycle fatigue

Stress components for LCF can be obtained from beam theory with known K factors or by hot spot FEanalysis for welded details or local FE analysis for free plate edges according to Sec6 The calculated stressranges for LCF should be corrected by using a plasticity correction factor in order to employ S-N curvesinstead of a strain-cycle curve

2 Number of static design cycles and fraction in different loadingconditionsThe number of design cycles depends on the ship type trade and design fatigue life TDF The design cycles inTable 1 should be used unless otherwise described and is referring to TD = 25 years

Table 1 Design cycles for low cycle fatigue nLCF

Ship type Design cycle nLCF

Tankers over 120 000 TDW 500

Tankers below 120 000 TDW 600

Chemical tankers 1000

App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000

Over Panamax bulk carriers 500

Panamax bulk carriers 800

Handymax bulk carriers about 45 000 TDW or smaller 1000

Shuttle tankers 1200

For vessels to be operated with frequent loading and unloading cycles the design cycle may be increasedbut need not be greater than 1 500 cycles for shuttle tankers chemical tankers and Handymax bulk carriersand 1 000 cycles for the other vessel types respectively

The fraction αk of load combination at sea is given in Table 2

Table 2 Fraction of load combination at sea for low cycle fatigue

Ship type Fraction of loadcombinations αk

Full load-Ballast L1 Alternate LCs L2

Tankers over 120 000 TDW 090 010

Tankers below 120 000 TDW 085 015

Chemical tankers 080 020

LNG carriers 100 000

LPG carriers 085 015

Over Panamax bulk carriers 090 010

Panamax bulk carriers and smaller 085 015

Ore carriers 085 015

Shuttle tankers 100 000

3 Loading conditions

31 Combination of static stressLoad conditions should be selected to obtain the static stress for each load condition and the stress rangefrom combination of those Figure 2 and Figure 3 show possible loading and unloading scenarios of a vesselduring a round trip The following two stress ranges should be taken into account at the design stage

App

endi

x H


DNV GL AS

Other possible load combinations eg full load to alternate or ballast to alternate can be disregarded

Figure 2 Operation scenarios full load - ballast

Figure 3 Operation scenarios ballast - full load ndash alternate load conditions

App

endi

x H


DNV GL AS

The static hot spot surface stress range or local stress range for free plate edge for low cycle fatigue shouldbe obtained from a combination of load conditions as shown in Table 3 Table 4 and Table 5

Table 3 Load combination for calculation of low cycle fatigue stress range ΔσLCF

Location Load conditions

Tankers withcentreline

longitudinalbulkhead

Tankers withtwo longitudinal

bulkheads

Vessels withoutlongitudinal bulkhead

Longitudinal flangeconnections ) Full load -ballast |σLC1-σLC2| |σLC7-σLC8| |σLC13-σLC14|

Web stiffener on top oflongitudinal stiffener Full load -ballast |σLC1-σLC2| |σLC7-σLC8| |σLC13-σLC14|

Transverse members weldedto longitudinals in waterballast tanks ie webstiffener cut out lug plate

Full load -ballast |σLC1-σLC2| |σLC7-σLC8| |σLC13-σLC14|

Lower and upper hopperknuckles lower and upperchamfers 1)


Full load -ballast |σLC1-σLC2| |σLC7-σLC8| |σLC13-σLC14|Horizontal stringer at innerside longitudinal bulkhead 1)

Alternate load |σLC3-σLC4| |σLC9-σLC10| |σLC15-σLC16|

Full load -ballast |σLC1-σLC2| |σLC7-σLC8| |σLC13-σLC14|Girder connection totransverse bulkhead innerbottom to lower stool innerbottom to cofferdam bulkhead1)

Alternate LCs 1-2 |σLC3-σLC4| |σLC9-σLC10| |σLC15-σLC16|

1) hull girder stress should be added to the local bending stress for the corresponding load condition in the trim andstability booklet

The load conditions in Table 4 may be applied to vessels with a centreline bulkhead Normal ballast conditionshall be used for ballast condition Actual draft for the alternate conditions Tact shall be obtained from theloading manual

Table 4 Load conditions for vessels with a centreline bulkhead

Load case Stress component Midship section view Plan view

LC1 Full load TSCσ LC 1

App

endi

x H


DNV GL AS


LC2 Ballast TBALσ LC 2

LC3 Alternate 1Tact σ LC 3



App

endi

x H


DNV GL AS



The load conditions in Table 5 may be applied to vessels with two longitudinal bulkheads As the loadcombination between load conditions LC5 and LC6 and between load conditions LC11 and LC12 is notcommon the combinations are neglected at the design stage

Table 5 Load conditions for vessels with two longitudinal bulkheads



LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12

The load conditions in Table 6 may be applied to vessels without longitudinal bulkhead eg LNG carriersbulk carriers ore carriers etc

Table 6 Load conditions for vessels without longitudinal bulkhead


LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS

LC14 Ballast Tballσ LC 14

LC15 Tact σ LC 15

LC16 Tact σ LC 16

32 Pressure loadsStresses contributed from both external pressure from sea water and internal pressure from both cargoand ballast water should be considered The calculation and establishment of external and internal dynamicpressures on 10-2 probability level of exceedance are given by the rulesThe static external pressure is calculated based on the draft as specified in Table 4 to Table 6 and accordingto the loading manual The static internal pressure is calculated based on actual filling height and cargodensity which can be obtained from the loading manual and the rules The static internal pressure should beestablished according to the loading conditions defined in Table 4 to Table 6

4 Simplified calculations of stresses

41 Hot spot stress range due to wave loadingWhere the principal stress direction for LCF is the same as for HCF the following procedure may be usedThe maximum expected stress range from the wave loading in the period between the loading and unloadingoperations should be added to the low cycle stress range before the fatigue damage from LCF is calculated

The hot spot stress range or local stress at free edge plate Δσj in Nmm2 at a probability of exceedance ofnLCF108 from the wave loading can be calculated as

App

endi

x H


DNV GL AS

where

ΔσHCFj = High cycle fatigue stress range for hot spot or local stress at free plate edge corresponding to 10-2

probability level for the loading condition j

n0 = number of cycles 108

42 Hot spot stress range for low cycle fatigueThe static elastic hot spot stress range (or local stress for free plate edge) for the load combination k for LCFcalculations is the difference between the hot spot stress (or local stress for free plate edge) components forloading condition i and j

Where

= static hot spot stress range for the load combination k given in Table 3

= static hot spot stress amplitude for loading condition i

= static hot spot stress amplitude for loading condition j

43 Combined hot spot stress rangeThe combined stress range for LCF representing a peak to peak stress due to loading and unloading andwave loading is given as

where

Δσi = Wave stress range at probability of nLCF108 for loading condition i

Δσj = Wave stress range at probability of nLCF108 for loading condition j

The effective pseudo stress range for calculation of LCF damage for the loading combination k can beobtained as

App

endi

x H


DNV GL AS

where

λn= Non-linearity correction factorke= Plasticity correction factor

= 10 for

= for

= Factor due to stress redistribution

= 10 if

= 09 for VLA-E if

= 08 for VL A-E32 or VL A-E36 steel if

σf = yield stressCoefficients for the plasticity correction factor a and b are given in Table 7

Table 7 Plasticity correction factor ke

Stress range

VL A-E steela = 116b = 0524

VL A-E32 and VL A-E36 steelsa = 10b = 053

The combined stress ranges should be derived from the linear elastic analysis The hot spot stress range (orlocal stress range at free plate edge) contributing to LCF is large and implies local yielding Thus a correctionof the elastic stress range is needed in order to derive a stress range that is representative for the actualstrain range taking the non-linear material behaviour into account

44 Plasticity correction factorThe plasticity correction factor can be obtained from an actual cyclic stress-strain curve and Neubers rules ornon-linear FE analysis as shown in Figure 4

App

endi

x H


DNV GL AS

where

= Elastic hot spot stress (or local stress at free plate edge) in Nmm2 obtained from linear elastic FEanalysis or a beam theory

= Pseudo linear elastic hot spot (or local stress at free plate edge) in Nmm2

=

For more complex structural connections only part of the region around the hot spot (or local stress area atthe free plate edge) will be yielding when subjected to large dynamic loads This can be accounted for by afactor accounting for redistribution of stress and strain Based on non-linear FE analysis of actual connectionsa redistribution factor may be introduced

In order to obtain the plasticity correction factor a cyclic stress-strain curve for materials should be obtainedfrom tests

If the cyclic stress-strain relation is combined with the Neubers rule the Neubers formula is given using theRamberg-Osgood relation as

where

K = stress concentration factorσhs = the actual stress in the hot spot (or local stress at free plate edge) in Nmm2 to be based on

non-linear FE analysis (hence it should not be based on nominal stress from linear FE times a Kfactor but taken from a curve as illustrated in Figure 4)

εhs = the actual strain in the hot spot (or local stress at free plated edge)E = Youngrsquos modulus in Nmm2

n Krsquo = material coefficients

K depends on the magnitude of the load and the sharpness of the notch Coefficients n and K are given inTable 8 for different steel grades used for derivation of the plasticity correction factors

Neuberrsquos rule is widely used to obtain the plasticity correction factor since it may give conservative results Ifthe plane strain behaviour is relevant the Glinka rule may be used for derivation of the plasticity correctionfactor instead of the Neuberrsquos rule

App

endi

x H


DNV GL AS

Table 8 Material properties for cyclic stress-strain curves

Material VL A-E VL A-E32 VL A-E36

Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115

Figure 4 Definition of stresses and strains

45 Total elastic stress componentThe elastic static hot spot stress amplitude due to static hull girder and pressure loads [42] and the elasticdynamic hot spot stress amplitude at nLCF108 probability level due to wave actions during loading andunloading [41] are established by use of beam theory or FE analysis as listed in Table 9 The relevantpressure and girder loads are given in the rules The dynamic stress components due to local and globalloads should be combined to a total dynamic stress range

Table 9 Stress models for elastic hotspot stress at LCF vulnerable location

Location Stress approach

Longitudinal flange connection Prescriptive stress analysis according to Sec4

Web stiffener on top of longitudinal stiffener Semi-nominal stress model according to DNVCN342 Plus - Extended fatigue analysis of shipdetails

Transverse members welded to longitudinals in water ballast tanksie cut-outs lug plate

Semi-nominal stress model according to DNVCN342 Plus - Extended fatigue analysis of shipdetails

App

endi

x H


DNV GL AS


Lower and upper hopper knuckles lower and upper chamfers Hot spot or local stress model according to Sec6

Horizontal stringer at inner side longitudinal bulkhead Hot spot or local stress model according to Sec6

Girder connection to transverse bulkhead inner bottom to lowerstool inner bottom to cofferdam bulkhead

Hot spot or local stress model according to Sec6

5 Fatigue damage calculations for low cycle fatigueA one-slope S-N curve (upper part of two slope S-N curves) for LCF strength is given as follows

Nk = Number of cycles to failure for LCF stress range

= Effective stress range in Nmm2 for the loading combination k

The basic S-N curve for low cycle fatigue assessment is given in Table 10 This design S-N curve is applicableto both welded joints and free plate edges for the LCF region The damage due to the LCF is calculated as

where

nLC = Total number of design load condition

αk = Fraction of load combinations see Table 2

If a non-linear FE analysis is carried out directly the effective pseudo-elastic hot spot stress amplitude can beobtained by multiplying the Youngrsquos modulus by the calculated notch strain amplitude

Table 10 Low cycle fatigue S-N curve

102 le N lt 104

Materiallog K2 m

Welded joints amp free plate edge 12164 30

6 Correction of the damage or stressNet or gross scantlings t should be used according to the rules for the specific ship type with the consistentscantlings approach factor fc when based on FE analysisS-N curve in air should be used for the entire design fatigue lifeThe thickness effect is not accounted for when evaluating the LCF damage

App

endi

x H


DNV GL AS

No mean stress effect should be considered for free plate edges and welded details for the LCF damageNo environmental reduction factor fe should be considered for evaluation of the LCF damageBenefit of weld improvement methods like grinding hammer-peening and TIG-dressing should not beappliedThe fabrication tolerances given in AppA are assumed applicable

7 Combined fatigue damage due to high and low cycle fatigueWhen DLCF ge 025 the combined damage ratio due to HCF and LCF is compared to the acceptance criteriaas

where

DHCF = Damage due to HCF for the design life = 25 years

DLCF = Damage due to LCF based on the design cycles not to be taken greater than the maximum design cycles in[2]

Note that the HCF damage contribution to the combined fatigue damage should be based on 25 years even ifan extended fatigue design life is required for the HCF calculationsFor DLCF lt 025 the fatigue damage is represented by the HCF damage with the acceptance criteria givenas

Figure 5 shows the requirements for the combined fatigue damages

Figure 5 The combined fatigue criteria

App

endi

x H


DNV GL AS

8 Example of applicationAn example of LCF strength assessment of a VLCC is illustrated Figure 6 shows a hot spot to be checked atan inner bottom longitudinal

Figure 6 Hot spot to be checked

The design condition and scantlings are given in Table 11

Table 11 Location to be checked

Item Requirements Remark

Design cycle nLCF 600 cycles From Table 1

Dimension of longitudinal 645 x 12 + 175 x 20 mm (T)VL A-E32 steel

Net or gross scantlingdepending on the rules for thespecific ship type

From FE analysis the assumed hot spot stress components are obtained at HS1 and given in Table 12

Table 12 Hot spot stress components at HS1 Nmm2

Stress components Full load Ballast

Hot spot stress due to still water vertical bending moment -1269 1269

Hot spot stress due to local bending of stiffener -270 3110

Total static hot spot stress σis -3969 4379

Dynamic stress range at 10-2 probability level σiHCF 3235 503

Dynamic stress range due to wave actions Δσi 875 1352

Thus the stress range for LCF is obtained as given in Table 13

App

endi

x H


DNV GL AS

Table 13 Combined stress range for low cycle fatigue strength assessment Nmm2

Stress component Full load-Ballast

Static hot spot stress range for low cycle fatigue DsLCFk 4379 ndash (-3969) = 8348

Combined stress range DsCombk 8348+ 05 (875+1352) = 9462

Figure 7 shows the hot spot stress components from loading and unloading and wave loading

Figure 7 Hot spot stress components Nmm2

The fatigue damage due to LCF is calculated in Table 14

Table 14 Low cycle fatigue strength assessment

Full load-Ballast

Plasticity correction ke 10middot9462middot10-3+053 = 148

Effective pseudo stress range Nmm2 08middot148middot9462 = 1 1203

The number of cycles Nk 1012164-3Log11203 = 1 038

Damage ratio due to HCF DHCF 024

Damage ratio due to LCF DLCF

Since the LCF damage is greater than 025 a combined damage due to HCF and LCF is calculated as

App

endi

x H


DNV GL AS

The current detail is acceptable in view of the combined fatigue due to LCF and HCF

App

endi

x I


DNV GL AS

APPENDIX I WAVE INDUCED VIBRATIONS FOR BLUNT VESSELS

1 IntroductionThe target vessels for this appendix are blunt vessels as oil tankers bulk carriers and ore carriers with blockcoefficient in the order of 08 or moreIn fatigue assessment of ship structures the waves induce

mdash quasi static stresses in the ship structure referred to as wave stressmdash dynamic vibrations of the hull girder referred to as vibration stress

The vibration stress comes from springing (resonance) and whipping (transient) and their relativeimportance depends on design (flexibility and shape) loading condition (ballast and cargo) and wavecondition (speed sea state and heading)Springing is caused by linear and non-linear excitation where the encounter frequency or the sum of twoencounter frequencies coincides with the natural frequency of the hull girder Whipping is caused by non-linear excitation like wave impact or slamming in the bow flare stem flare bottom and stern area The twophenomena occur more or less continuously and simultaneously and may be difficult to distinguish due tolow vibration damping Therefore they are commonly referred to as wave induced vibrations from a fatigueconsequence point of viewThe governing vibration shape as shown in Figure 1 is the two-node vertical mode which is associated withthe lowest natural vibration frequency It is most easily excited and gives the largest vertical bending stressamidships

Figure 1 Upper is 2-node vertical bending moment and lower plot is associated vertical bendingmoment distribution for a normalised homogeneous ship

The damping is an important parameter which affects the vibration level of springing and the decay of thespringing and whipping in lack of excitation The damping is low for the governing vibration modesThe period of the vibration stress (025-3 seconds) is an order of magnitude lower than the periods of thewave stress (5-20 seconds) The vibration stress combined with the wave stress makes up a broad bandprocess and Rainflow counting is then the recognised approach to establish the loading history First thefatigue damage is calculated for the total stress (wave stress + vibration stress) which defines the totaldamage Secondly the fatigue damage is estimated for the wave stress referred to as the wave damage Thedifference between the total and wave damage makes up the vibration damage In practise it is the vibrationon top of the wave frequency loading that makes up the significant part of the vibration damage for ocean

App

endi

x I


DNV GL AS

going vessels The vibration damage has been of comparable magnitude as the wave damage for all full scalemeasurements and model tests that have been assessed but the relative magnitude depends on ship typeand trade

2 How to include the effect of vibrationThe effect of vibration should be applied to the simplified fatigue assessmentIt is assumed that the vibration stress is put on top of the wave stress from the vertical wave bendingmoment only The vertical wave stress to be used in the fatigue assessment is then adjusted by

Where

fvib = vibration factor which represents a correction of the wave stress consistent with the additional vibrationdamage (from whipping and springing) for the intended design area eg North Atlantic or World Wide Thecorrection factor assumes all wave headings of equal probability

σvw = denotes the wave stress σv from vertical bending moment which can be derived from the first component ofthe global hull girder bending stress in Sec4 [4]

The vibration factor fvib is given as fvib = fvibj where the subscript j refers to the loading condition j It can beestimated as

Where

Dwj = Wave damage in loading condition j

Dvibi = Vibration damage in loading condition j

m = Inverse slope of the SN-curve consistent with the estimate of the damages

The ship specific estimate of fvib can replace the same factor in the rules (RU SHIP Pt3 Ch4 Sec4 [311])

3 Vibration factor for blunt vesselsBlunt vessels are vessels with high block coefficients like ore carriers Empirical vibration factor is basedmainly on full scale measurement data with some correction from model tests for ships with block coefficientof about 08 and with design speeds of about 15 knots It is supposed to cover a size range from about 170to 350 meters in length The correction factor is estimated as

Where

App

endi

x I


DNV GL AS

B = Moulded hull breadth (m)

CB = Block coefficient at scantling draught

Lpp = Length between perpendiculars (m)

Z = Hull girder section modulus gross scantlings (m3)

m = 3 for welded material and 4 for base material assumed protected from corrosive environment

R = route factor

= 0937 for North Atlantic operation

= 10 for World Wide operation

Ti = forward draught in m in loading condition i i = c for cargo and i = b for ballast It is recommended to use draftrelated to heavy ballast condition or gale ballast draft

V = contract speed at design draft at 85 MCR and 20 sea margin in knots

If the contract speed Vd is specified at another x MCR and y sea margin it can be converted by thefollowing formula (simplistic)

For hatch corners (where m=4 is relevant) the vibration factor is only applied as a correction to wave stressfrom vertical bending moment

4 Application of the vibration factor fvibIn the most simplified approach useful for early design the fvib is replacing fvib in the rules related to thedownscaling of the wave bending moment to a 10-2 probability level of exceedance as shown in RU SHIP Pt3Ch4 Sec4 [311] This implies that the fatigue vertical bending moment is multiplied with fvibIt is however convenient to utilize the directly calculated vertical wave bending moment towards the forwardand aft part of the cargo area A reduction factor fd(x) can be established based on the directly calculatedfatigue moment distribution along the hull

App

endi

x I


DNV GL AS

Where

Mwv-LCj = Directly calculated wave bending moment for position x from AP for loading condition j The moment istaken out at a probability level of exceedance of 10-2

The fd(x) is applied as a reduction factor to the maximum vertical wave bending moment amidships Itthereby replaces the moment distribution factor fm in the rules (however fm may be used in the software andthen this need to be accounted for) The fd(x) ie the normalised bending moment distributions for ballastand cargo condition is illustrated Figure 2

Figure 2 Reduction factor fd(x) as a function of length in ballast and cargo condition (anexample)

The stress from the vertical bending moment is now calculated as

Where denotes the wave stress as determined from the first component of the global hull girder

bending stress in [4] but where fm = 10 consistent with the maximum moments amidships

5 Effect of the tradeStandard design trades are the North Atlantic or World Wide If another specific trade is specified by theowneryard in agreement with the Society the environmental factor fe can be estimated This can also be

App

endi

x I


DNV GL AS

useful if another wave source than global wave statistics are to be used or small vessels are intended forharsh design trade since the relation between World Wide and North Atlantic may differ from large vesselsThe environmental factor fe is established for the vessel or for a similar vessel by component stochasticfatigue analysis For convenience it is sufficient to consider only the vertical wave bending moment Thefatigue damage is calculated for the North Atlantic and the actual scatter diagram for a life time of 25 yearsand for both loading conditions (as the route specific scatter diagram may also differ for the two loadingconditions due to sailing restrictions) This gives the following damages

Dba = Damage for ballast condition in actual trade

Dca = Damage for cargo condition in actual trade

Dbww = Damage for ballast condition in North Atlantic

Dcww = Damage for cargo condition in North Atlantic

The part time in the different conditions is denoted αj where j = b for ballast and j = c for cargo The parttime is taken from the rules or as specified The environmental factor is estimated in the following principleway

Where m can be taken as 3 and represents most of the fatigue sensitive welded details

6 Model tests procedureWhile empirical relations are useful in early design the most accurate way of establishing fvib is by modeltests This assumes that the state-of-the-art procedure is followedAnother benefit of model tests is that the vibration factor fvib can be estimated along the vessel while theempirical relations assume a constant value in the cargo areaIn agreement with the owneryard the model test procedure should be submitted for approval

7 Numerical calculation procedureNumerical analysis can replace model tests It is regarded less accurate than model tests but can still bea useful alternative in order to estimate ship specific fvib factors This assumes that the state-of-the-areprocedure is followed and that the numerical tool is tuned for the particular vessel designThe state-of-the-art hydroelastic tools today are regarded more useful for container vessels with dominatingwhipping impacts than blunt vessels with dominating non-linear springing excitationIn agreement with the owneryard the model test procedure should be submitted for approval

8 Full scale measurementsThe uncertainties in the encountered wave environment in different trades are considerable in model testsand numerical analysis Full scale measurements of similar ships on similar trades can also be basis forestimation of fvib factors for sister vessels or future similar designsThe full scale measurements for obtaining useful data in decision support should preferably be carried outusing approved hull monitoring system according to the rules for Hull monitoring systems (class notationHMON) Systems approved according to 2005 revision or later include fatigue and extreme loading with andwithout the effect of vibration included but it should also be ensured that measured raw and statistical datais stored and can be submitted to shore for further assessment

App

endi

x I


DNV GL AS

In agreement with the owneryard the documentation should be submitted for approval

App

endi

x J


DNV GL AS

APPENDIX J CONSIDERATIONS OF SPECIAL DETAILS

1 Crane foundations and foundations for heavy top side loads

11 IntroductionCrane foundations permanent foundations for top side equipment and preliminary foundations for cargoare integrated into to the hull structure The foundations may be exposed to high loads from the top sideequipment and cargo but also to hull girder wave loads When these foundations are located close tomidship the nominal stress from the hull girder wave loads in way of the foundation may be significant Ifthe foundation is not designed with due consideration to the risk of fatigue damage cracks may appear inoperationIt may be sufficient to consider the fatigue damage from the top side loads and the hull girder wave loadsseparately but for some designs it is necessary to assess the combined effect of the two main load effectsFor cranes the specific operation of the crane may also need to be considered and for optimized designseven low cycle fatigue assessment may be necessary This is however regarded as outside the scope of thehull girder fatigue assessment and subject to requirements for lifting appliancesA brief guidance is given in the following

12 Fatigue assessment based on wave induced loadsThe support details exposed to longitudinal (axial) stresses from the global hull girder loads should beconsideredFor support details located in the shadow zone from global hull girder loads aft or forward of large deckopenings fatigue assessment based on hull girder wave loads may be neglectedUnless the support details are covered by standard details in AppA the support details should be assessedby FE analysis using the hot spot stress approachFor the foundations nominal stress level for the top side loads at the deck level should be estimated basedon inertia loads caused by ship motions When the largest dynamic vertical nominal stress exceed

mdash the dynamic nominal longitudinal stress at the deck corner even though the foundation is located in theshadow zone the foundation should be assessed by FE analysis due to the top side loads

mdash half the dynamic nominal longitudinal stress at the deck level where the foundation is located theinteraction between the longitudinal hull girder stress and the vertical stress from the top side loadsshould be considered The loads should be assumed to be simultaneously acting

If the vertical dynamic nominal stress is less or equal to half of the longitudinal stress from the hull girderwave loads at deck corner the foundation can be assessed as if the top side loads are not presentBoth the top side loads and the hull girder wave loads should be based on the same probability levelof exceedance (10-2) If the top side loads can vary due to different possible arrangements the mostunfavourable realistic arrangement should serve as basis for the estimate of the vertical nominal stress

DNV GLDriven by our purpose of safeguarding life property and the environment DNV GL enablesorganizations to advance the safety and sustainability of their business We provide classification andtechnical assurance along with software and independent expert advisory services to the maritimeoil and gas and energy industries We also provide certification services to customers across a widerange of industries Operating in more than 100 countries our 16 000 professionals are dedicated tohelping our customers make the world safer smarter and greener

SAFER SMARTER GREENER

CONTENTS

Changes ndash current

Section 1 General

1 Introduction


21 Symbols

22 Abbreviations

23 Coordinate system and sign conventions

3 Application

31 General

32 Calculated and observed fatigue strength

33 Vibrations

34 Low cycle fatigue


41 Stress types used for fatigue assessment

42 Methods for stress calculation

43 Selection of prescriptive or direct methods for calculation of stresses


51 General

52 Crack failure modes

53 Assessment using local and nominal stress approach

54 Assessment using hot spot stress approach

55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions

Section 2 Fatigue Capacity

1 Introduction

2 S-N curves

21 General

22 S-N curves and detail classification

23 S-N curves for in-air environment

24 S-N curves for stress along the welds

25 Nominal stress S-N curves

26 Other steel types


31 General

32 Fatigue assessment in corrosive environment

4 Mean stress effect

41 Base material and free plate edges

42 Welded joints

5 Thickness effect

51 General

52 Effective thickness of welds

53 Thickness effect of welds in prescriptive analysis of longitudinal end connections

54 Thickness effect of welds in FE analysis

55 Thickness effect of base material

6 Material factor

Section 3 Fatigue strength representation

1 Introduction


21 General

22 Scantlings approach factor fc

23 Environmental factor fe

24 Post-weld treatment



31 Fatigue damage accumulation with Palmgren - Minerrsquos rule

32 Time in air and corrosive environment

33 Annual fatigue damage

34 The combined fatigue damage

35 The total fatigue damage from multiple loading conditions

36 Fatigue life

4 Permissible stress range

5 Permissible stress concentration factor and required FAT class

Section 4 Prescriptive fatigue strength assessment

1 Introduction

11 Stress approach for longitudinal end connections

12 Calculation procedure for longitudinal end connections

13 Definition of stress components

14 Calculation of stress components

2 Wave induced loads


31 Hot spot stress or nominal stress

32 Prescriptive fatigue stress range

33 Stress range

34 Mean stress


41 Stress due to wave induced hull girder loads

42 Hull girder vibrations

43 Stress due to still water hull girder bending moment


51 General


61 Wave induced stiffener bending stress

62 Still water stiffener bending stress


71 General

72 Relative deflection definition

73 Sign convention

74 Estimate of relative deflection

75 Prescriptive stress due to relative deflection derived by FE analysis

76 Stress due to relative displacement in still water

77 Consideration of reduced local stiffener bending stress


81 Longitudinal stiffener end connections

82 Other connection types and overlapped connections

Section 5 Direct Fatigue strength assessment

1 General

11 Definition

12 Stress transfer functions

13 Assumptions

14 Hydrodynamic theory


21 Fatigue damage calculations

3 Component stochastic analysis

31 Load components

32 From load to stress transfer functions

33 Stress factors per unit load

34 Splash zone correction

35 Double hull and relative deflection stress


41 Introduction

42 Assessment of local details

Section 6 Finite Element Analysis

1 Introduction

11 Thickness

12 Application

13 Stress to be used from finite element analysis

14 Stress field at a welded detail

15 Fatigue stress range based on EDW approach

16 Fatigue stress range based on direct calculations

17 Hot spot S-N curve


21 General

22 FE models

23 Load application

24 Mesh size

25 4-node or 8-node elements

26 Base material free plate edges

27 Example Hatch corners and hatch coaming end bracket


31 General

32 Read out point at t2

33 Derivation of stress at read out point t2

34 Reduced hot spot stress for highly localized stress peaks with plate bending


41 Procedure for analysis of hot spot stress at weld toe of welded details

42 Procedure for analysis of local stress of base material at free plate edge

5 Procedure for analysis of hot spot stress of web-stiffened cruciform joints

51 General

52 Calculation of hot spot stress at the flange

53 Calculation of hot spot stress in way of the web

6 Procedure for analysis of hot spot stress of simple cruciform joints

61 Limitation to hot spot stress approach

62 Correction of hot spot stress approach for simple cruciform joints

Section 7 Improvement of Fatigue Life by Fabrication

1 Introduction

11 General

2 Grinding

21 Weld toe burr grinding

22 Flush grinding of butt welds

23 Fatigue life improvement factor for burr grinding

Section 8 Detail design standard

1 Introduction

2 Detail design standard for longitudinal end connections without top stiffener

21 Design standard A - slots with and without lug plate

22 Design standard B - collar plate

23 Equivalent design of longitudinal end connections without top stiffener

3 Detail design standard for different ship types

Section 9 References

1 References

Appendix A Stress Concentration Factors and FAT classes

1 General

11 Definition of stress concentration factors

12 Determination of stress concentration factors

13 Definition of FAT class

14 Basis


21 Tabulated K factors

22 Soft toe of web stiffener and backing bracket

23 Unsymmetrical stiffener flange

3 Butt welds

4 Flange connections


51 Attachments welded to a plate or stiffener

52 Termination of stiffeners

53 Doubling plates

6 Simple cruciform joints

7 Scallops and block joints

8 Lower hopper knuckles

9 Pipe connections

10 Holes with edge reinforcements

11 Rounded rectangular elliptical and oval holes

Appendix B Fatigue capacity

1 Failure criterion inherent in the S-N curves

2 Other S-N curves

21 S-N curves for base material of high strength steel above 500Nmm2

22 Qualification of new S-N curves based on fatigue test data

3 Effect of fabrication tolerances

4 Design chart for fillet and partial penetration welds

Appendix C Fatigue damage calculations and fatigue stress design tables

1 Weibull long term stress distribution


21 Practical damage calculation with one slope S-N curve

22 Closed form damage estimate for one-slope S-N curves and Weibull distribution

23 Closed form damage estimate for two-slope S-N curves and Weibull distribution

24 Closed form damage estimate for one slope S-N curve and short term Rayleigh distribution

25 Closed form damage estimate for two-slope S-N curve and short term Rayleigh distribution

3 Maximum allowable stress range

4 Guidance for omission of fatigue analysis

Appendix D Prescriptive fatigue strength assessment


11 Prescriptive double hull bending stress at transverse bulkhead

12 Prescriptive wave induced double side bending stress at the transverse bulkhead

13 Prescriptive still water double side bending stress at the transverse bulkhead


21 Prescriptive wave induced relative deflection for double hull vessels

22 Prescriptive relative deflection for double hull vessels due to static pressure


31 Wave induced plate bending stress

32 Still water plate bending stress

Appendix E Finite element analysis

1 FE modelling

11 Mesh size

12 Solid elements


21 Extrapolation from t2 and 3t2

22 Derivation of effective hot spot stress considering principal stress directions


31 Derivation of stress at read out points t2 and 3t2


41 Bent hopper tank knuckles


51 Introduction

52 Procedure

53 Screening fatigue criteria

54 Stress read out procedure for bracket toe

55 Stress read out procedure for knuckle detail

56 Read out point stress

6 Derivation of stress concentration factors for alternative end connection designs

7 Verification of analysis methodology

Appendix F Workmanship and link to analysis procedures

1 Tolerance limits

2 Weld profiling by machining and grinding

3 TIG dressing

4 Hammer peening

5 Improvement factors for TIG dressing and Hammer peening

Appendix G Uncertainties in Fatigue Life Predictions

1 Introduction

2 Probability of fatigue failure

3 Size of defects inherent in S-N data and requirements to non-destructive examination

Appendix H Low cycle fatigue

1 General

2 Number of static design cycles and fraction in different loading conditions


31 Combination of static stress

32 Pressure loads


41 Hot spot stress range due to wave loading

42 Hot spot stress range for low cycle fatigue

43 Combined hot spot stress range

44 Plasticity correction factor

45 Total elastic stress component

5 Fatigue damage calculations for low cycle fatigue

6 Correction of the damage or stress

7 Combined fatigue damage due to high and low cycle fatigue

8 Example of application

Appendix I Wave induced vibrations for blunt vessels

1 Introduction

2 How to include the effect of vibration

3 Vibration factor for blunt vessels

4 Application of the vibration factor fvib

5 Effect of the trade

6 Model tests procedure

7 Numerical calculation procedure

8 Full scale measurements

Appendix J Considerations of special details


11 Introduction

12 Fatigue assessment based on wave induced loads

FOREWORD

DNV GL class guidelines contain methods technical requirements principles and acceptancecriteria related to classed objects as referred to from the rules

copy DNV GL AS

Any comments may be sent by e-mail to rulesdnvglcom

If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of DNV GL then DNV GL shallpay compensation to such person for his proved direct loss or damage However the compensation shall not exceed an amount equal to tentimes the fee charged for the service in question provided that the maximum compensation shall never exceed USD 2 million

In this provision DNV GL shall mean DNV GL AS its direct and indirect owners as well as all its affiliates subsidiaries directors officersemployees agents and any other acting on behalf of DNV GL

Cha

nges

- c

urre

nt


DNV GL AS



Con

tent

s


DNV GL AS

CONTENTS








7 Definitions 29


Con

tent

s


DNV GL AS











Con

tent

s


DNV GL AS












Con

tent

s


DNV GL AS










Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS
















Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS









Sec

tion

1


DNV GL AS

SECTION 1 GENERAL




Table 1 Symbol list

Symbol Meaning Unit








Γ( )

Γ( )


-



mm








Sec

tion

1


DNV GL AS

Symbol Meaning Unit












ηWηldηbd

ηSηlsηbs



-



q Roll angle rad










σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)


Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit


Nmm2












Nmm2






σmean i(j)pX

σmean i(j)pY


Nmm2


Nmm2







Nmm2


Sec

tion

1


DNV GL AS

Symbol Meaning Unit









Nmm2



Γ( )

Γ( )


-









mm



-



mm





Sec

tion

1


DNV GL AS

Symbol Meaning Unit






-
















IFwd-n50

IAft-n50








(Nmm2)m


(Nmm2)m


Sec

tion

1


DNV GL AS

Symbol Meaning Unit


-








ℓbdgℓbdg Fwd

ℓbdg Aft






-


Nmm2














-



Sec

tion

1


DNV GL AS

Symbol Meaning Unit

































Sec

tion

1


DNV GL AS

Symbol Meaning Unit











xexeFwd

xeAft


m



m





m



m


Zeff-n50ZAft-n50

ZFwd-n50


cm3


mm




Sec

tion

1

Class guideline mdash DNVGL-CG-0129 Edition October 2015 0Fatigue assessment of ship structures

DNV GL AS


FE Finite Element



CG Class Guideline









WF Wave Frequency


NA North Atlantic

WW World Wide












AF Axial Force



Sec

tion

1


DNV GL AS




3 Application



doubling plates

Sec

tion

1


DNV GL AS











Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS













Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS





Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS


Sec

tion

1


DNV GL AS


Sec

tion

2


DNV GL AS





2 S-N curves








Sec

tion

2


DNV GL AS


log K2


where




= 020




Sec

tion

2


DNV GL AS


S-NCurve


cycles (FAT)


(knuckle) Δσq


(107) cycles

N le 107 N gt 107


B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5



Sec

tion

2


DNV GL AS



FAT classNmm2

1


B1

160

2


B

150

3


B2 140

4


C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5


C1 112

6


C2

100



Sec

tion

2


DNV GL AS


KpFATNmm2


072125



080112



090100





Sec

tion

2


DNV GL AS

KpFATNmm2


11380









Sec

tion

2


DNV GL AS





Sec

tion

2


DNV GL AS


5 Thickness effect


for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where









02



01

Sec

tion

2


DNV GL AS





01



00

Sec

tion

2


DNV GL AS




0100



00



00



Sec

tion

2


DNV GL AS









Sec

tion

2


DNV GL AS




Sec

tion

3


DNV GL AS













Sec

tion

3


DNV GL AS









where





Sec

tion

3


DNV GL AS





Sec

tion

3


DNV GL AS

where






Sec

tion

3


DNV GL AS

where












where





Sec

tion

3


DNV GL AS

where






for

for

for

where

and where


Sec

tion

3


DNV GL AS






fN ==

=

=



Sec

tion

3


DNV GL AS



where






Sec

tion

4


DNV GL AS


1 Introduction








Sec

tion

4


DNV GL AS



Sec

tion

4


DNV GL AS











Sec

tion

4


DNV GL AS






where



where






Sec

tion

4


DNV GL AS





where




where






Sec

tion

4


DNV GL AS

where















Sec

tion

4


DNV GL AS

where









where




Sec

tion

4


DNV GL AS





Sec

tion

4


DNV GL AS

where






= -1 otherwise








Sec

tion

4


DNV GL AS


Sec

tion

4


DNV GL AS


where






η = -1 otherwise









Sec

tion

4


DNV GL AS


where




σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)



Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50





δFwdik(j)

δAftik(j)





Sec

tion

4


DNV GL AS




Sec

tion

5


DNV GL AS


1 General










Sec

tion

5


DNV GL AS







Sec

tion

5


DNV GL AS




accelerations



Sec

tion

5


DNV GL AS



where
















Sec

tion

5


DNV GL AS



where





Sec

tion

5


DNV GL AS





Sec

tion

5


DNV GL AS






Sec

tion

5


DNV GL AS



Sec

tion

6


DNV GL AS


1 Introduction




Type Description






Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS




where




Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)



where















Sec

tion

6


DNV GL AS



where



where



fmean1(j)fmean2(j)





where



Sec

tion

6


DNV GL AS













Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS




Sec

tion

6


DNV GL AS









Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS



where





Sec

tion

6


DNV GL AS



where



where







Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS

where



where





Sec

tion

6


DNV GL AS


where










Sec

tion

6


DNV GL AS



where



where


Sec

tion

6


DNV GL AS

where





Sec

tion

6


DNV GL AS



where




Sec

tion

7


DNV GL AS


1 Introduction


2 Grinding


Sec

tion

7


DNV GL AS



Sec

tion

7


DNV GL AS





Sec

tion

7


DNV GL AS







Sec

tion

8


DNV GL AS















Sec

tion

8


DNV GL AS



Design standard A

1 2

3 4




Sec

tion

8


DNV GL AS


Design standard A

1 2










Sec

tion

8


DNV GL AS



Sec

tion

8


DNV GL AS





bottom




Sec

tion

8


DNV GL AS


Sec

tion

9


DNV GL AS
























Sec

tion

9


DNV GL AS















App

endi

x A


DNV GL AS


1 General




where








App

endi

x A


DNV GL AS






Point A Point BNo


Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250


160 for dgt250

114 for dle 150124


127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114





App

endi

x A


DNV GL AS




where




App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


where






App

endi

x A


DNV GL AS




fA = Reduction



= 10 in other cases



App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS




200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2


01317hstf- 35726

(tw-gr + 0006hstf- 03139)2




1



080 112

2



100 90

App

endi

x A


DNV GL AS


3



113 80

Quality level C2)


125 72

4



113 80

Quality level C2)


125 72

5



113 80

App

endi

x A


DNV GL AS


6




127 71

Quality level C2)


142 63

7





248 36

Quality level C2)


278 32

8




248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9








10


where



11


Where







90104middotKm3)







App

endi

x A


DNV GL AS


12



Kme = 10Kmα = 10





1


147 61

App

endi

x A


DNV GL AS


2




127

143

71

63

3



18

20

50

45

App

endi

x A


DNV GL AS


4



10

127

18

90

71

50

5


where


147 61

6


180 50

App

endi

x A


DNV GL AS


7




B-C2 150-100

8




10 90

9




09 100

App

endi

x A


DNV GL AS


10



Connection length

Nominal stress

143 63

11




113 80

12


slope 15

slope 13

slope 12


113

127

143

80

71

63

App

endi

x A


DNV GL AS


13


40 23





1



t gt 25 mm and





113

127

119

80

71

76

2



113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3


-Flat bar

-Bulb profile

-Angle profile


160

160

180

56

56

50

4


-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5



160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6


Non-load carrying

Load carrying

113

20

80

45




1



90K

App

endi

x A


DNV GL AS


2




90K




1


50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d


-reinforced ends or




120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2


50 lt d le 100

100 lt d le 150

150 lt d le 300


127

132

143

161

71

68

63

56





1


Cruciform joint

Tee-joint

127

113

71

80

2


Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3



σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4


t le 25 mm

t gt 25 mm

105

119

86

76

5




App

endi

x A


DNV GL AS









1


Point A

Point B

24

127

375

71

2


Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3


Point A

Point B

117

127

77

71

4


Point A

Point B

117

127

77

71




App

endi

x A


DNV GL AS



1


70 13

2


45 20

3


25 36

App

endi

x A


DNV GL AS




1


For tubular section



161

180

56

50

2


For tubular section



200

225

45

40

3


d le 50 mm

d gt 50 mm


127

143

71

63



App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS




00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS








App

endi

x A


DNV GL AS


App

endi

x B


DNV GL AS



2 Other S-N curves



App

endi

x B


DNV GL AS



App

endi

x B


DNV GL AS




App

endi

x B


DNV GL AS


App

endi

x C


DNV GL AS



where





where






App

endi

x C


DNV GL AS

where








where







cycles

App

endi

x C


DNV GL AS


ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306


where







App

endi

x C


DNV GL AS

where












App

endi

x C


DNV GL AS




=









060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS







060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS


In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45


App

endi

x C


DNV GL AS



App

endi

x D


DNV GL AS





App

endi

x D


DNV GL AS



where









App

endi

x D


DNV GL AS










where





App

endi

x D


DNV GL AS



where









where





App

endi

x D


DNV GL AS

where







ia ib = IaS Ibℓ



App

endi

x D


DNV GL AS



App

endi

x D


DNV GL AS

where







where





App

endi

x D


DNV GL AS





App

endi

x E


DNV GL AS


1 FE modelling








App

endi

x E


DNV GL AS





App

endi

x E


DNV GL AS













Solid elements


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS

a)

b)


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS



where




App

endi

x E


DNV GL AS


App

endi

x E


DNV GL AS













App

endi

x E


DNV GL AS











166

Bulk carrier






App

endi

x E


DNV GL AS




App

endi

x E


DNV GL AS



App

endi

x E


DNV GL AS












App

endi

x E


DNV GL AS







App

endi

x F


DNV GL AS











15 mmt2







App

endi

x F


DNV GL AS



AppA Table 8




tt10 max 15 mm






where




App

endi

x F


DNV GL AS









App

endi

x F


DNV GL AS










App

endi

x G


DNV GL AS




App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS



App

endi

x H


DNV GL AS











App

endi

x H


DNV GL AS









App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000





















App

endi

x H


DNV GL AS




App

endi

x H


DNV GL AS







bulkheads

















App

endi

x H


DNV GL AS






App

endi

x H


DNV GL AS







LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12




LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS


LC15 Tact σ LC 15

LC16 Tact σ LC 16





App

endi

x H


DNV GL AS

where





Where





where




App

endi

x H


DNV GL AS

where


= 10 for

= for


= 10 if

= 09 for VLA-E if




Stress range





App

endi

x H


DNV GL AS

where



=




where







App

endi

x H


DNV GL AS



Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115









App

endi

x H


DNV GL AS










where





102 le N lt 104

Materiallog K2 m



App

endi

x H


DNV GL AS



where






App

endi

x H


DNV GL AS


















App

endi

x H


DNV GL AS









Full load-Ballast







App

endi

x H


DNV GL AS


App

endi

x I


DNV GL AS







App

endi

x I


DNV GL AS



Where




Where






Where

App

endi

x I


DNV GL AS






R = route factor








App

endi

x I


DNV GL AS

Where








App

endi

x I


DNV GL AS











App

endi

x I


DNV GL AS


App

endi

x J


DNV GL AS










CONTENTS


Section 1 General

1 Introduction


21 Symbols

22 Abbreviations


3 Application

31 General


33 Vibrations







51 General




55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions


1 Introduction

2 S-N curves

21 General







31 General




42 Welded joints

5 Thickness effect

51 General





6 Material factor


1 Introduction


21 General











36 Fatigue life




1 Introduction









33 Stress range

34 Mean stress






51 General





71 General


73 Sign convention









1 General

11 Definition


13 Assumptions





31 Load components






41 Introduction



1 Introduction

11 Thickness

12 Application







21 General

22 FE models

23 Load application

24 Mesh size





31 General








51 General







1 Introduction

11 General

2 Grinding





1 Introduction







1 References


1 General




14 Basis





3 Butt welds





53 Doubling plates




9 Pipe connections





2 Other S-N curves



























1 FE modelling

11 Mesh size

12 Solid elements









51 Introduction

52 Procedure








1 Tolerance limits


3 TIG dressing

4 Hammer peening



1 Introduction




1 General




32 Pressure loads












1 Introduction










11 Introduction


Cha

nges

- c

urre

nt

Class guideline mdash DNVGL-CG-0129 Edition October 2015 Fatigue assessment of ship structures

DNV GL AS



Con

tent

s


DNV GL AS

CONTENTS








7 Definitions 29


Con

tent

s


DNV GL AS











Con

tent

s


DNV GL AS












Con

tent

s


DNV GL AS










Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS
















Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS









Sec

tion

1


DNV GL AS

SECTION 1 GENERAL




Table 1 Symbol list

Symbol Meaning Unit








Γ( )

Γ( )


-



mm








Sec

tion

1


DNV GL AS

Symbol Meaning Unit












ηWηldηbd

ηSηlsηbs



-



q Roll angle rad










σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)


Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit


Nmm2












Nmm2






σmean i(j)pX

σmean i(j)pY


Nmm2


Nmm2







Nmm2


Sec

tion

1


DNV GL AS

Symbol Meaning Unit









Nmm2



Γ( )

Γ( )


-









mm



-



mm





Sec

tion

1


DNV GL AS

Symbol Meaning Unit






-
















IFwd-n50

IAft-n50








(Nmm2)m


(Nmm2)m


Sec

tion

1


DNV GL AS

Symbol Meaning Unit


-








ℓbdgℓbdg Fwd

ℓbdg Aft






-


Nmm2














-



Sec

tion

1


DNV GL AS

Symbol Meaning Unit

































Sec

tion

1


DNV GL AS

Symbol Meaning Unit











xexeFwd

xeAft


m



m





m



m


Zeff-n50ZAft-n50

ZFwd-n50


cm3


mm




Sec

tion

1


DNV GL AS


FE Finite Element



CG Class Guideline









WF Wave Frequency


NA North Atlantic

WW World Wide












AF Axial Force



Sec

tion

1


DNV GL AS




3 Application



doubling plates

Sec

tion

1


DNV GL AS











Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS













Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS





Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS


Sec

tion

1


DNV GL AS


Sec

tion

2


DNV GL AS





2 S-N curves








Sec

tion

2


DNV GL AS


log K2


where




= 020




Sec

tion

2


DNV GL AS


S-NCurve


cycles (FAT)


(knuckle) Δσq


(107) cycles

N le 107 N gt 107


B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5



Sec

tion

2


DNV GL AS



FAT classNmm2

1


B1

160

2


B

150

3


B2 140

4


C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5


C1 112

6


C2

100



Sec

tion

2


DNV GL AS


KpFATNmm2


072125



080112



090100





Sec

tion

2


DNV GL AS

KpFATNmm2


11380









Sec

tion

2


DNV GL AS





Sec

tion

2


DNV GL AS


5 Thickness effect


for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where









02



01

Sec

tion

2


DNV GL AS





01



00

Sec

tion

2


DNV GL AS




0100



00



00



Sec

tion

2


DNV GL AS









Sec

tion

2


DNV GL AS




Sec

tion

3


DNV GL AS













Sec

tion

3


DNV GL AS









where





Sec

tion

3


DNV GL AS





Sec

tion

3


DNV GL AS

where






Sec

tion

3


DNV GL AS

where












where





Sec

tion

3


DNV GL AS

where






for

for

for

where

and where


Sec

tion

3


DNV GL AS






fN ==

=

=



Sec

tion

3


DNV GL AS



where






Sec

tion

4


DNV GL AS


1 Introduction








Sec

tion

4


DNV GL AS



Sec

tion

4


DNV GL AS











Sec

tion

4


DNV GL AS






where



where






Sec

tion

4


DNV GL AS





where




where






Sec

tion

4


DNV GL AS

where















Sec

tion

4


DNV GL AS

where









where




Sec

tion

4


DNV GL AS





Sec

tion

4


DNV GL AS

where






= -1 otherwise








Sec

tion

4


DNV GL AS


Sec

tion

4


DNV GL AS


where






η = -1 otherwise









Sec

tion

4


DNV GL AS


where




σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)



Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50





δFwdik(j)

δAftik(j)





Sec

tion

4


DNV GL AS




Sec

tion

5


DNV GL AS


1 General










Sec

tion

5


DNV GL AS







Sec

tion

5


DNV GL AS




accelerations



Sec

tion

5


DNV GL AS



where
















Sec

tion

5


DNV GL AS



where





Sec

tion

5


DNV GL AS





Sec

tion

5


DNV GL AS






Sec

tion

5


DNV GL AS



Sec

tion

6


DNV GL AS


1 Introduction




Type Description






Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS




where




Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)



where















Sec

tion

6


DNV GL AS



where



where



fmean1(j)fmean2(j)





where



Sec

tion

6


DNV GL AS













Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS




Sec

tion

6


DNV GL AS









Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS



where





Sec

tion

6


DNV GL AS



where



where







Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS

where



where





Sec

tion

6


DNV GL AS


where










Sec

tion

6


DNV GL AS



where



where


Sec

tion

6


DNV GL AS

where





Sec

tion

6


DNV GL AS



where




Sec

tion

7


DNV GL AS


1 Introduction


2 Grinding


Sec

tion

7


DNV GL AS



Sec

tion

7


DNV GL AS





Sec

tion

7


DNV GL AS







Sec

tion

8


DNV GL AS















Sec

tion

8


DNV GL AS



Design standard A

1 2

3 4




Sec

tion

8


DNV GL AS


Design standard A

1 2










Sec

tion

8


DNV GL AS



Sec

tion

8


DNV GL AS





bottom




Sec

tion

8


DNV GL AS


Sec

tion

9


DNV GL AS
























Sec

tion

9


DNV GL AS















App

endi

x A


DNV GL AS


1 General




where








App

endi

x A


DNV GL AS






Point A Point BNo


Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250


160 for dgt250

114 for dle 150124


127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114





App

endi

x A


DNV GL AS




where




App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


where






App

endi

x A


DNV GL AS




fA = Reduction



= 10 in other cases



App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS




200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2


01317hstf- 35726

(tw-gr + 0006hstf- 03139)2




1



080 112

2



100 90

App

endi

x A


DNV GL AS


3



113 80

Quality level C2)


125 72

4



113 80

Quality level C2)


125 72

5



113 80

App

endi

x A


DNV GL AS


6




127 71

Quality level C2)


142 63

7





248 36

Quality level C2)


278 32

8




248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9








10


where



11


Where







90104middotKm3)







App

endi

x A


DNV GL AS


12



Kme = 10Kmα = 10





1


147 61

App

endi

x A


DNV GL AS


2




127

143

71

63

3



18

20

50

45

App

endi

x A


DNV GL AS


4



10

127

18

90

71

50

5


where


147 61

6


180 50

App

endi

x A


DNV GL AS


7




B-C2 150-100

8




10 90

9




09 100

App

endi

x A


DNV GL AS


10



Connection length

Nominal stress

143 63

11




113 80

12


slope 15

slope 13

slope 12


113

127

143

80

71

63

App

endi

x A


DNV GL AS


13


40 23





1



t gt 25 mm and





113

127

119

80

71

76

2



113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3


-Flat bar

-Bulb profile

-Angle profile


160

160

180

56

56

50

4


-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5



160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6


Non-load carrying

Load carrying

113

20

80

45




1



90K

App

endi

x A


DNV GL AS


2




90K




1


50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d


-reinforced ends or




120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2


50 lt d le 100

100 lt d le 150

150 lt d le 300


127

132

143

161

71

68

63

56





1


Cruciform joint

Tee-joint

127

113

71

80

2


Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3



σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4


t le 25 mm

t gt 25 mm

105

119

86

76

5




App

endi

x A


DNV GL AS









1


Point A

Point B

24

127

375

71

2


Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3


Point A

Point B

117

127

77

71

4


Point A

Point B

117

127

77

71




App

endi

x A


DNV GL AS



1


70 13

2


45 20

3


25 36

App

endi

x A


DNV GL AS




1


For tubular section



161

180

56

50

2


For tubular section



200

225

45

40

3


d le 50 mm

d gt 50 mm


127

143

71

63



App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS




00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS








App

endi

x A


DNV GL AS


App

endi

x B


DNV GL AS



2 Other S-N curves



App

endi

x B


DNV GL AS



App

endi

x B


DNV GL AS




App

endi

x B


DNV GL AS


App

endi

x C


DNV GL AS



where





where






App

endi

x C


DNV GL AS

where








where







cycles

App

endi

x C


DNV GL AS


ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306


where







App

endi

x C


DNV GL AS

where












App

endi

x C


DNV GL AS




=









060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS







060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS


In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45


App

endi

x C


DNV GL AS



App

endi

x D


DNV GL AS





App

endi

x D


DNV GL AS



where









App

endi

x D


DNV GL AS










where





App

endi

x D


DNV GL AS



where









where





App

endi

x D


DNV GL AS

where







ia ib = IaS Ibℓ



App

endi

x D


DNV GL AS



App

endi

x D


DNV GL AS

where







where





App

endi

x D


DNV GL AS





App

endi

x E


DNV GL AS


1 FE modelling








App

endi

x E


DNV GL AS





App

endi

x E


DNV GL AS













Solid elements


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS

a)

b)


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS



where




App

endi

x E


DNV GL AS


App

endi

x E


DNV GL AS













App

endi

x E


DNV GL AS











166

Bulk carrier






App

endi

x E


DNV GL AS




App

endi

x E


DNV GL AS



App

endi

x E


DNV GL AS












App

endi

x E


DNV GL AS







App

endi

x F


DNV GL AS











15 mmt2







App

endi

x F


DNV GL AS



AppA Table 8




tt10 max 15 mm






where




App

endi

x F


DNV GL AS









App

endi

x F


DNV GL AS










App

endi

x G


DNV GL AS




App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS



App

endi

x H


DNV GL AS











App

endi

x H


DNV GL AS









App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000





















App

endi

x H


DNV GL AS




App

endi

x H


DNV GL AS







bulkheads

















App

endi

x H


DNV GL AS






App

endi

x H


DNV GL AS







LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12




LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS


LC15 Tact σ LC 15

LC16 Tact σ LC 16





App

endi

x H


DNV GL AS

where





Where





where




App

endi

x H


DNV GL AS

where


= 10 for

= for


= 10 if

= 09 for VLA-E if




Stress range





App

endi

x H


DNV GL AS

where



=




where







App

endi

x H


DNV GL AS



Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115









App

endi

x H


DNV GL AS










where





102 le N lt 104

Materiallog K2 m



App

endi

x H


DNV GL AS



where






App

endi

x H


DNV GL AS


















App

endi

x H


DNV GL AS









Full load-Ballast







App

endi

x H


DNV GL AS


App

endi

x I


DNV GL AS







App

endi

x I


DNV GL AS



Where




Where






Where

App

endi

x I


DNV GL AS






R = route factor








App

endi

x I


DNV GL AS

Where








App

endi

x I


DNV GL AS











App

endi

x I


DNV GL AS


App

endi

x J


DNV GL AS










CONTENTS


Section 1 General

1 Introduction


21 Symbols

22 Abbreviations


3 Application

31 General


33 Vibrations







51 General




55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions


1 Introduction

2 S-N curves

21 General







31 General




42 Welded joints

5 Thickness effect

51 General





6 Material factor


1 Introduction


21 General











36 Fatigue life




1 Introduction









33 Stress range

34 Mean stress






51 General





71 General


73 Sign convention









1 General

11 Definition


13 Assumptions





31 Load components






41 Introduction



1 Introduction

11 Thickness

12 Application







21 General

22 FE models

23 Load application

24 Mesh size





31 General








51 General







1 Introduction

11 General

2 Grinding





1 Introduction







1 References


1 General




14 Basis





3 Butt welds





53 Doubling plates




9 Pipe connections





2 Other S-N curves



























1 FE modelling

11 Mesh size

12 Solid elements









51 Introduction

52 Procedure








1 Tolerance limits


3 TIG dressing

4 Hammer peening



1 Introduction




1 General




32 Pressure loads












1 Introduction










11 Introduction


Con

tent

s


DNV GL AS

CONTENTS








7 Definitions 29


Con

tent

s


DNV GL AS











Con

tent

s


DNV GL AS












Con

tent

s


DNV GL AS










Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS
















Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS









Sec

tion

1


DNV GL AS

SECTION 1 GENERAL




Table 1 Symbol list

Symbol Meaning Unit








Γ( )

Γ( )


-



mm








Sec

tion

1


DNV GL AS

Symbol Meaning Unit












ηWηldηbd

ηSηlsηbs



-



q Roll angle rad










σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)


Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit


Nmm2












Nmm2






σmean i(j)pX

σmean i(j)pY


Nmm2


Nmm2







Nmm2


Sec

tion

1


DNV GL AS

Symbol Meaning Unit









Nmm2



Γ( )

Γ( )


-









mm



-



mm





Sec

tion

1


DNV GL AS

Symbol Meaning Unit






-
















IFwd-n50

IAft-n50








(Nmm2)m


(Nmm2)m


Sec

tion

1


DNV GL AS

Symbol Meaning Unit


-








ℓbdgℓbdg Fwd

ℓbdg Aft






-


Nmm2














-



Sec

tion

1


DNV GL AS

Symbol Meaning Unit

































Sec

tion

1


DNV GL AS

Symbol Meaning Unit











xexeFwd

xeAft


m



m





m



m


Zeff-n50ZAft-n50

ZFwd-n50


cm3


mm




Sec

tion

1


DNV GL AS


FE Finite Element



CG Class Guideline









WF Wave Frequency


NA North Atlantic

WW World Wide












AF Axial Force



Sec

tion

1


DNV GL AS




3 Application



doubling plates

Sec

tion

1


DNV GL AS











Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS













Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS





Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS


Sec

tion

1


DNV GL AS


Sec

tion

2


DNV GL AS





2 S-N curves








Sec

tion

2


DNV GL AS


log K2


where




= 020




Sec

tion

2


DNV GL AS


S-NCurve


cycles (FAT)


(knuckle) Δσq


(107) cycles

N le 107 N gt 107


B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5



Sec

tion

2


DNV GL AS



FAT classNmm2

1


B1

160

2


B

150

3


B2 140

4


C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5


C1 112

6


C2

100



Sec

tion

2


DNV GL AS


KpFATNmm2


072125



080112



090100





Sec

tion

2


DNV GL AS

KpFATNmm2


11380









Sec

tion

2


DNV GL AS





Sec

tion

2


DNV GL AS


5 Thickness effect


for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where









02



01

Sec

tion

2


DNV GL AS





01



00

Sec

tion

2


DNV GL AS




0100



00



00



Sec

tion

2


DNV GL AS









Sec

tion

2


DNV GL AS




Sec

tion

3


DNV GL AS













Sec

tion

3


DNV GL AS









where





Sec

tion

3


DNV GL AS





Sec

tion

3


DNV GL AS

where






Sec

tion

3


DNV GL AS

where












where





Sec

tion

3


DNV GL AS

where






for

for

for

where

and where


Sec

tion

3


DNV GL AS






fN ==

=

=



Sec

tion

3


DNV GL AS



where






Sec

tion

4


DNV GL AS


1 Introduction








Sec

tion

4


DNV GL AS



Sec

tion

4


DNV GL AS











Sec

tion

4


DNV GL AS






where



where






Sec

tion

4


DNV GL AS





where




where






Sec

tion

4


DNV GL AS

where















Sec

tion

4


DNV GL AS

where









where




Sec

tion

4


DNV GL AS





Sec

tion

4


DNV GL AS

where






= -1 otherwise








Sec

tion

4


DNV GL AS


Sec

tion

4


DNV GL AS


where






η = -1 otherwise









Sec

tion

4


DNV GL AS


where




σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)



Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50





δFwdik(j)

δAftik(j)





Sec

tion

4


DNV GL AS




Sec

tion

5


DNV GL AS


1 General










Sec

tion

5


DNV GL AS







Sec

tion

5


DNV GL AS




accelerations



Sec

tion

5


DNV GL AS



where
















Sec

tion

5


DNV GL AS



where





Sec

tion

5


DNV GL AS





Sec

tion

5


DNV GL AS






Sec

tion

5


DNV GL AS



Sec

tion

6


DNV GL AS


1 Introduction




Type Description






Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS




where




Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)



where















Sec

tion

6


DNV GL AS



where



where



fmean1(j)fmean2(j)





where



Sec

tion

6


DNV GL AS













Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS




Sec

tion

6


DNV GL AS









Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS



where





Sec

tion

6


DNV GL AS



where



where







Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS

where



where





Sec

tion

6


DNV GL AS


where










Sec

tion

6


DNV GL AS



where



where


Sec

tion

6


DNV GL AS

where





Sec

tion

6


DNV GL AS



where




Sec

tion

7


DNV GL AS


1 Introduction


2 Grinding


Sec

tion

7


DNV GL AS



Sec

tion

7


DNV GL AS





Sec

tion

7


DNV GL AS







Sec

tion

8


DNV GL AS















Sec

tion

8


DNV GL AS



Design standard A

1 2

3 4




Sec

tion

8


DNV GL AS


Design standard A

1 2










Sec

tion

8


DNV GL AS



Sec

tion

8


DNV GL AS





bottom




Sec

tion

8


DNV GL AS


Sec

tion

9


DNV GL AS
























Sec

tion

9


DNV GL AS















App

endi

x A


DNV GL AS


1 General




where








App

endi

x A


DNV GL AS






Point A Point BNo


Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250


160 for dgt250

114 for dle 150124


127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114





App

endi

x A


DNV GL AS




where




App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


where






App

endi

x A


DNV GL AS




fA = Reduction



= 10 in other cases



App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS




200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2


01317hstf- 35726

(tw-gr + 0006hstf- 03139)2




1



080 112

2



100 90

App

endi

x A


DNV GL AS


3



113 80

Quality level C2)


125 72

4



113 80

Quality level C2)


125 72

5



113 80

App

endi

x A


DNV GL AS


6




127 71

Quality level C2)


142 63

7





248 36

Quality level C2)


278 32

8




248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9








10


where



11


Where







90104middotKm3)







App

endi

x A


DNV GL AS


12



Kme = 10Kmα = 10





1


147 61

App

endi

x A


DNV GL AS


2




127

143

71

63

3



18

20

50

45

App

endi

x A


DNV GL AS


4



10

127

18

90

71

50

5


where


147 61

6


180 50

App

endi

x A


DNV GL AS


7




B-C2 150-100

8




10 90

9




09 100

App

endi

x A


DNV GL AS


10



Connection length

Nominal stress

143 63

11




113 80

12


slope 15

slope 13

slope 12


113

127

143

80

71

63

App

endi

x A


DNV GL AS


13


40 23





1



t gt 25 mm and





113

127

119

80

71

76

2



113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3


-Flat bar

-Bulb profile

-Angle profile


160

160

180

56

56

50

4


-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5



160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6


Non-load carrying

Load carrying

113

20

80

45




1



90K

App

endi

x A


DNV GL AS


2




90K




1


50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d


-reinforced ends or




120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2


50 lt d le 100

100 lt d le 150

150 lt d le 300


127

132

143

161

71

68

63

56





1


Cruciform joint

Tee-joint

127

113

71

80

2


Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3



σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4


t le 25 mm

t gt 25 mm

105

119

86

76

5




App

endi

x A


DNV GL AS









1


Point A

Point B

24

127

375

71

2


Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3


Point A

Point B

117

127

77

71

4


Point A

Point B

117

127

77

71




App

endi

x A


DNV GL AS



1


70 13

2


45 20

3


25 36

App

endi

x A


DNV GL AS




1


For tubular section



161

180

56

50

2


For tubular section



200

225

45

40

3


d le 50 mm

d gt 50 mm


127

143

71

63



App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS




00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS








App

endi

x A


DNV GL AS


App

endi

x B


DNV GL AS



2 Other S-N curves



App

endi

x B


DNV GL AS



App

endi

x B


DNV GL AS




App

endi

x B


DNV GL AS


App

endi

x C


DNV GL AS



where





where






App

endi

x C


DNV GL AS

where








where







cycles

App

endi

x C


DNV GL AS


ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306


where







App

endi

x C


DNV GL AS

where












App

endi

x C


DNV GL AS




=









060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS







060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS


In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45


App

endi

x C


DNV GL AS



App

endi

x D


DNV GL AS





App

endi

x D


DNV GL AS



where









App

endi

x D


DNV GL AS










where





App

endi

x D


DNV GL AS



where









where





App

endi

x D


DNV GL AS

where







ia ib = IaS Ibℓ



App

endi

x D


DNV GL AS



App

endi

x D


DNV GL AS

where







where





App

endi

x D


DNV GL AS





App

endi

x E


DNV GL AS


1 FE modelling








App

endi

x E


DNV GL AS





App

endi

x E


DNV GL AS













Solid elements


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS

a)

b)


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS



where




App

endi

x E


DNV GL AS


App

endi

x E


DNV GL AS













App

endi

x E


DNV GL AS











166

Bulk carrier






App

endi

x E


DNV GL AS




App

endi

x E


DNV GL AS



App

endi

x E


DNV GL AS












App

endi

x E


DNV GL AS







App

endi

x F


DNV GL AS











15 mmt2







App

endi

x F


DNV GL AS



AppA Table 8




tt10 max 15 mm






where




App

endi

x F


DNV GL AS









App

endi

x F


DNV GL AS










App

endi

x G


DNV GL AS




App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS



App

endi

x H


DNV GL AS











App

endi

x H


DNV GL AS









App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000





















App

endi

x H


DNV GL AS




App

endi

x H


DNV GL AS







bulkheads

















App

endi

x H


DNV GL AS






App

endi

x H


DNV GL AS







LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12




LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS


LC15 Tact σ LC 15

LC16 Tact σ LC 16





App

endi

x H


DNV GL AS

where





Where





where




App

endi

x H


DNV GL AS

where


= 10 for

= for


= 10 if

= 09 for VLA-E if




Stress range





App

endi

x H


DNV GL AS

where



=




where







App

endi

x H


DNV GL AS



Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115









App

endi

x H


DNV GL AS










where





102 le N lt 104

Materiallog K2 m



App

endi

x H


DNV GL AS



where






App

endi

x H


DNV GL AS


















App

endi

x H


DNV GL AS









Full load-Ballast







App

endi

x H


DNV GL AS


App

endi

x I


DNV GL AS







App

endi

x I


DNV GL AS



Where




Where






Where

App

endi

x I


DNV GL AS






R = route factor








App

endi

x I


DNV GL AS

Where








App

endi

x I


DNV GL AS











App

endi

x I


DNV GL AS


App

endi

x J


DNV GL AS










CONTENTS


Section 1 General

1 Introduction


21 Symbols

22 Abbreviations


3 Application

31 General


33 Vibrations







51 General




55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions


1 Introduction

2 S-N curves

21 General







31 General




42 Welded joints

5 Thickness effect

51 General





6 Material factor


1 Introduction


21 General











36 Fatigue life




1 Introduction









33 Stress range

34 Mean stress






51 General





71 General


73 Sign convention









1 General

11 Definition


13 Assumptions





31 Load components






41 Introduction



1 Introduction

11 Thickness

12 Application







21 General

22 FE models

23 Load application

24 Mesh size





31 General








51 General







1 Introduction

11 General

2 Grinding





1 Introduction







1 References


1 General




14 Basis





3 Butt welds





53 Doubling plates




9 Pipe connections





2 Other S-N curves



























1 FE modelling

11 Mesh size

12 Solid elements









51 Introduction

52 Procedure








1 Tolerance limits


3 TIG dressing

4 Hammer peening



1 Introduction




1 General




32 Pressure loads












1 Introduction










11 Introduction


Con

tent

s


DNV GL AS











Con

tent

s


DNV GL AS












Con

tent

s


DNV GL AS










Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS
















Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS









Sec

tion

1


DNV GL AS

SECTION 1 GENERAL




Table 1 Symbol list

Symbol Meaning Unit








Γ( )

Γ( )


-



mm








Sec

tion

1


DNV GL AS

Symbol Meaning Unit












ηWηldηbd

ηSηlsηbs



-



q Roll angle rad










σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)


Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit


Nmm2












Nmm2






σmean i(j)pX

σmean i(j)pY


Nmm2


Nmm2







Nmm2


Sec

tion

1


DNV GL AS

Symbol Meaning Unit









Nmm2



Γ( )

Γ( )


-









mm



-



mm





Sec

tion

1


DNV GL AS

Symbol Meaning Unit






-
















IFwd-n50

IAft-n50








(Nmm2)m


(Nmm2)m


Sec

tion

1


DNV GL AS

Symbol Meaning Unit


-








ℓbdgℓbdg Fwd

ℓbdg Aft






-


Nmm2














-



Sec

tion

1


DNV GL AS

Symbol Meaning Unit

































Sec

tion

1


DNV GL AS

Symbol Meaning Unit











xexeFwd

xeAft


m



m





m



m


Zeff-n50ZAft-n50

ZFwd-n50


cm3


mm




Sec

tion

1


DNV GL AS


FE Finite Element



CG Class Guideline









WF Wave Frequency


NA North Atlantic

WW World Wide












AF Axial Force



Sec

tion

1


DNV GL AS




3 Application



doubling plates

Sec

tion

1


DNV GL AS











Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS













Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS





Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS


Sec

tion

1


DNV GL AS


Sec

tion

2


DNV GL AS





2 S-N curves








Sec

tion

2


DNV GL AS


log K2


where




= 020




Sec

tion

2


DNV GL AS


S-NCurve


cycles (FAT)


(knuckle) Δσq


(107) cycles

N le 107 N gt 107


B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5



Sec

tion

2


DNV GL AS



FAT classNmm2

1


B1

160

2


B

150

3


B2 140

4


C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5


C1 112

6


C2

100



Sec

tion

2


DNV GL AS


KpFATNmm2


072125



080112



090100





Sec

tion

2


DNV GL AS

KpFATNmm2


11380









Sec

tion

2


DNV GL AS





Sec

tion

2


DNV GL AS


5 Thickness effect


for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where









02



01

Sec

tion

2


DNV GL AS





01



00

Sec

tion

2


DNV GL AS




0100



00



00



Sec

tion

2


DNV GL AS









Sec

tion

2


DNV GL AS




Sec

tion

3


DNV GL AS













Sec

tion

3


DNV GL AS









where





Sec

tion

3


DNV GL AS





Sec

tion

3


DNV GL AS

where






Sec

tion

3


DNV GL AS

where












where





Sec

tion

3


DNV GL AS

where






for

for

for

where

and where


Sec

tion

3


DNV GL AS






fN ==

=

=



Sec

tion

3


DNV GL AS



where






Sec

tion

4


DNV GL AS


1 Introduction








Sec

tion

4


DNV GL AS



Sec

tion

4


DNV GL AS











Sec

tion

4


DNV GL AS






where



where






Sec

tion

4


DNV GL AS





where




where






Sec

tion

4


DNV GL AS

where















Sec

tion

4


DNV GL AS

where









where




Sec

tion

4


DNV GL AS





Sec

tion

4


DNV GL AS

where






= -1 otherwise








Sec

tion

4


DNV GL AS


Sec

tion

4


DNV GL AS


where






η = -1 otherwise









Sec

tion

4


DNV GL AS


where




σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)



Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50





δFwdik(j)

δAftik(j)





Sec

tion

4


DNV GL AS




Sec

tion

5


DNV GL AS


1 General










Sec

tion

5


DNV GL AS







Sec

tion

5


DNV GL AS




accelerations



Sec

tion

5


DNV GL AS



where
















Sec

tion

5


DNV GL AS



where





Sec

tion

5


DNV GL AS





Sec

tion

5


DNV GL AS






Sec

tion

5


DNV GL AS



Sec

tion

6


DNV GL AS


1 Introduction




Type Description






Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS




where




Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)



where















Sec

tion

6


DNV GL AS



where



where



fmean1(j)fmean2(j)





where



Sec

tion

6


DNV GL AS













Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS




Sec

tion

6


DNV GL AS









Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS



where





Sec

tion

6


DNV GL AS



where



where







Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS

where



where





Sec

tion

6


DNV GL AS


where










Sec

tion

6


DNV GL AS



where



where


Sec

tion

6


DNV GL AS

where





Sec

tion

6


DNV GL AS



where




Sec

tion

7


DNV GL AS


1 Introduction


2 Grinding


Sec

tion

7


DNV GL AS



Sec

tion

7


DNV GL AS





Sec

tion

7


DNV GL AS







Sec

tion

8


DNV GL AS















Sec

tion

8


DNV GL AS



Design standard A

1 2

3 4




Sec

tion

8


DNV GL AS


Design standard A

1 2










Sec

tion

8


DNV GL AS



Sec

tion

8


DNV GL AS





bottom




Sec

tion

8


DNV GL AS


Sec

tion

9


DNV GL AS
























Sec

tion

9


DNV GL AS















App

endi

x A


DNV GL AS


1 General




where








App

endi

x A


DNV GL AS






Point A Point BNo


Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250


160 for dgt250

114 for dle 150124


127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114





App

endi

x A


DNV GL AS




where




App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


where






App

endi

x A


DNV GL AS




fA = Reduction



= 10 in other cases



App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS




200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2


01317hstf- 35726

(tw-gr + 0006hstf- 03139)2




1



080 112

2



100 90

App

endi

x A


DNV GL AS


3



113 80

Quality level C2)


125 72

4



113 80

Quality level C2)


125 72

5



113 80

App

endi

x A


DNV GL AS


6




127 71

Quality level C2)


142 63

7





248 36

Quality level C2)


278 32

8




248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9








10


where



11


Where







90104middotKm3)







App

endi

x A


DNV GL AS


12



Kme = 10Kmα = 10





1


147 61

App

endi

x A


DNV GL AS


2




127

143

71

63

3



18

20

50

45

App

endi

x A


DNV GL AS


4



10

127

18

90

71

50

5


where


147 61

6


180 50

App

endi

x A


DNV GL AS


7




B-C2 150-100

8




10 90

9




09 100

App

endi

x A


DNV GL AS


10



Connection length

Nominal stress

143 63

11




113 80

12


slope 15

slope 13

slope 12


113

127

143

80

71

63

App

endi

x A


DNV GL AS


13


40 23





1



t gt 25 mm and





113

127

119

80

71

76

2



113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3


-Flat bar

-Bulb profile

-Angle profile


160

160

180

56

56

50

4


-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5



160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6


Non-load carrying

Load carrying

113

20

80

45




1



90K

App

endi

x A


DNV GL AS


2




90K




1


50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d


-reinforced ends or




120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2


50 lt d le 100

100 lt d le 150

150 lt d le 300


127

132

143

161

71

68

63

56





1


Cruciform joint

Tee-joint

127

113

71

80

2


Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3



σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4


t le 25 mm

t gt 25 mm

105

119

86

76

5




App

endi

x A


DNV GL AS









1


Point A

Point B

24

127

375

71

2


Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3


Point A

Point B

117

127

77

71

4


Point A

Point B

117

127

77

71




App

endi

x A


DNV GL AS



1


70 13

2


45 20

3


25 36

App

endi

x A


DNV GL AS




1


For tubular section



161

180

56

50

2


For tubular section



200

225

45

40

3


d le 50 mm

d gt 50 mm


127

143

71

63



App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS




00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS








App

endi

x A


DNV GL AS


App

endi

x B


DNV GL AS



2 Other S-N curves



App

endi

x B


DNV GL AS



App

endi

x B


DNV GL AS




App

endi

x B


DNV GL AS


App

endi

x C


DNV GL AS



where





where






App

endi

x C


DNV GL AS

where








where







cycles

App

endi

x C


DNV GL AS


ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306


where







App

endi

x C


DNV GL AS

where












App

endi

x C


DNV GL AS




=









060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS







060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS


In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45


App

endi

x C


DNV GL AS



App

endi

x D


DNV GL AS





App

endi

x D


DNV GL AS



where









App

endi

x D


DNV GL AS










where





App

endi

x D


DNV GL AS



where









where





App

endi

x D


DNV GL AS

where







ia ib = IaS Ibℓ



App

endi

x D


DNV GL AS



App

endi

x D


DNV GL AS

where







where





App

endi

x D


DNV GL AS





App

endi

x E


DNV GL AS


1 FE modelling








App

endi

x E


DNV GL AS





App

endi

x E


DNV GL AS













Solid elements


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS

a)

b)


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS



where




App

endi

x E


DNV GL AS


App

endi

x E


DNV GL AS













App

endi

x E


DNV GL AS











166

Bulk carrier






App

endi

x E


DNV GL AS




App

endi

x E


DNV GL AS



App

endi

x E


DNV GL AS












App

endi

x E


DNV GL AS







App

endi

x F


DNV GL AS











15 mmt2







App

endi

x F


DNV GL AS



AppA Table 8




tt10 max 15 mm






where




App

endi

x F


DNV GL AS









App

endi

x F


DNV GL AS










App

endi

x G


DNV GL AS




App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS



App

endi

x H


DNV GL AS











App

endi

x H


DNV GL AS









App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000





















App

endi

x H


DNV GL AS




App

endi

x H


DNV GL AS







bulkheads

















App

endi

x H


DNV GL AS






App

endi

x H


DNV GL AS







LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12




LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS


LC15 Tact σ LC 15

LC16 Tact σ LC 16





App

endi

x H


DNV GL AS

where





Where





where




App

endi

x H


DNV GL AS

where


= 10 for

= for


= 10 if

= 09 for VLA-E if




Stress range





App

endi

x H


DNV GL AS

where



=




where







App

endi

x H


DNV GL AS



Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115









App

endi

x H


DNV GL AS










where





102 le N lt 104

Materiallog K2 m



App

endi

x H


DNV GL AS



where






App

endi

x H


DNV GL AS


















App

endi

x H


DNV GL AS









Full load-Ballast







App

endi

x H


DNV GL AS


App

endi

x I


DNV GL AS







App

endi

x I


DNV GL AS



Where




Where






Where

App

endi

x I


DNV GL AS






R = route factor








App

endi

x I


DNV GL AS

Where








App

endi

x I


DNV GL AS











App

endi

x I


DNV GL AS


App

endi

x J


DNV GL AS










CONTENTS


Section 1 General

1 Introduction


21 Symbols

22 Abbreviations


3 Application

31 General


33 Vibrations







51 General




55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions


1 Introduction

2 S-N curves

21 General







31 General




42 Welded joints

5 Thickness effect

51 General





6 Material factor


1 Introduction


21 General











36 Fatigue life




1 Introduction









33 Stress range

34 Mean stress






51 General





71 General


73 Sign convention









1 General

11 Definition


13 Assumptions





31 Load components






41 Introduction



1 Introduction

11 Thickness

12 Application







21 General

22 FE models

23 Load application

24 Mesh size





31 General








51 General







1 Introduction

11 General

2 Grinding





1 Introduction







1 References


1 General




14 Basis





3 Butt welds





53 Doubling plates




9 Pipe connections





2 Other S-N curves



























1 FE modelling

11 Mesh size

12 Solid elements









51 Introduction

52 Procedure








1 Tolerance limits


3 TIG dressing

4 Hammer peening



1 Introduction




1 General




32 Pressure loads












1 Introduction










11 Introduction


Con

tent

s


DNV GL AS












Con

tent

s


DNV GL AS










Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS
















Con

tent

s


DNV GL AS















Con

tent

s


DNV GL AS









Sec

tion

1


DNV GL AS

SECTION 1 GENERAL




Table 1 Symbol list

Symbol Meaning Unit








Γ( )

Γ( )


-



mm








Sec

tion

1


DNV GL AS

Symbol Meaning Unit












ηWηldηbd

ηSηlsηbs



-



q Roll angle rad










σdFwd-aik(j)

σdFwd-fik(j)

σdAft-aik(j)

σdAft-fik(j)


Nmm2

Sec

tion

1


DNV GL AS

Symbol Meaning Unit


Nmm2












Nmm2






σmean i(j)pX

σmean i(j)pY


Nmm2


Nmm2







Nmm2


Sec

tion

1


DNV GL AS

Symbol Meaning Unit









Nmm2



Γ( )

Γ( )


-









mm



-



mm





Sec

tion

1


DNV GL AS

Symbol Meaning Unit






-
















IFwd-n50

IAft-n50








(Nmm2)m


(Nmm2)m


Sec

tion

1


DNV GL AS

Symbol Meaning Unit


-








ℓbdgℓbdg Fwd

ℓbdg Aft






-


Nmm2














-



Sec

tion

1


DNV GL AS

Symbol Meaning Unit

































Sec

tion

1


DNV GL AS

Symbol Meaning Unit











xexeFwd

xeAft


m



m





m



m


Zeff-n50ZAft-n50

ZFwd-n50


cm3


mm




Sec

tion

1


DNV GL AS


FE Finite Element



CG Class Guideline









WF Wave Frequency


NA North Atlantic

WW World Wide












AF Axial Force



Sec

tion

1


DNV GL AS




3 Application



doubling plates

Sec

tion

1


DNV GL AS











Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS













Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS





Sec

tion

1


DNV GL AS




Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS








Sec

tion

1


DNV GL AS






Sec

tion

1


DNV GL AS


Sec

tion

1


DNV GL AS


Sec

tion

2


DNV GL AS





2 S-N curves








Sec

tion

2


DNV GL AS


log K2


where




= 020




Sec

tion

2


DNV GL AS


S-NCurve


cycles (FAT)


(knuckle) Δσq


(107) cycles

N le 107 N gt 107


B1 160 10700 056 (049) 15118 4 19176 6

B 150 10031 060 (052) 15005 4 19008 6

B2 140 9362 064 (056) 14886 4 18828 6

C 125 7892 072 (067) 13640 35 17435 55

C1 112 7072 080 (074) 13473 35 17172 55

C2 100 6314 090 (083) 13301 35 16902 55

D 90 5263 100 (100) 12164 3 15606 5

E 80 4678 113 (113) 12010 3 15350 5

F 71 4152 127 (127) 11855 3 15091 5



Sec

tion

2


DNV GL AS



FAT classNmm2

1


B1

160

2


B

150

3


B2 140

4


C 125

Sec

tion

2


DNV GL AS


FAT classNmm2

5


C1 112

6


C2

100



Sec

tion

2


DNV GL AS


KpFATNmm2


072125



080112



090100





Sec

tion

2


DNV GL AS

KpFATNmm2


11380









Sec

tion

2


DNV GL AS





Sec

tion

2


DNV GL AS


5 Thickness effect


for teff gt 25 mm

Sec

tion

2


DNV GL AS

for teff le 25 mm

where









02



01

Sec

tion

2


DNV GL AS





01



00

Sec

tion

2


DNV GL AS




0100



00



00



Sec

tion

2


DNV GL AS









Sec

tion

2


DNV GL AS




Sec

tion

3


DNV GL AS













Sec

tion

3


DNV GL AS









where





Sec

tion

3


DNV GL AS





Sec

tion

3


DNV GL AS

where






Sec

tion

3


DNV GL AS

where












where





Sec

tion

3


DNV GL AS

where






for

for

for

where

and where


Sec

tion

3


DNV GL AS






fN ==

=

=



Sec

tion

3


DNV GL AS



where






Sec

tion

4


DNV GL AS


1 Introduction








Sec

tion

4


DNV GL AS



Sec

tion

4


DNV GL AS











Sec

tion

4


DNV GL AS






where



where






Sec

tion

4


DNV GL AS





where




where






Sec

tion

4


DNV GL AS

where















Sec

tion

4


DNV GL AS

where









where




Sec

tion

4


DNV GL AS





Sec

tion

4


DNV GL AS

where






= -1 otherwise








Sec

tion

4


DNV GL AS


Sec

tion

4


DNV GL AS


where






η = -1 otherwise









Sec

tion

4


DNV GL AS


where




σdFwd-aik(j)

σdAft-aik(j)

σdFwd-fik(j)

σdAft-fik(j)



Sec

tion

4


DNV GL AS

ZFwd-n50 ZAft-

n50





δFwdik(j)

δAftik(j)





Sec

tion

4


DNV GL AS




Sec

tion

5


DNV GL AS


1 General










Sec

tion

5


DNV GL AS







Sec

tion

5


DNV GL AS




accelerations



Sec

tion

5


DNV GL AS



where
















Sec

tion

5


DNV GL AS



where





Sec

tion

5


DNV GL AS





Sec

tion

5


DNV GL AS






Sec

tion

5


DNV GL AS



Sec

tion

6


DNV GL AS


1 Introduction




Type Description






Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS




where




Sec

tion

6


DNV GL AS

where

ΔσBSi1(j)

ΔσBSi2(j)



where















Sec

tion

6


DNV GL AS



where



where



fmean1(j)fmean2(j)





where



Sec

tion

6


DNV GL AS













Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS




Sec

tion

6


DNV GL AS









Sec

tion

6


DNV GL AS






Sec

tion

6


DNV GL AS



where





Sec

tion

6


DNV GL AS



where



where







Sec

tion

6


DNV GL AS





Sec

tion

6


DNV GL AS

where



where





Sec

tion

6


DNV GL AS


where










Sec

tion

6


DNV GL AS



where



where


Sec

tion

6


DNV GL AS

where





Sec

tion

6


DNV GL AS



where




Sec

tion

7


DNV GL AS


1 Introduction


2 Grinding


Sec

tion

7


DNV GL AS



Sec

tion

7


DNV GL AS





Sec

tion

7


DNV GL AS







Sec

tion

8


DNV GL AS















Sec

tion

8


DNV GL AS



Design standard A

1 2

3 4




Sec

tion

8


DNV GL AS


Design standard A

1 2










Sec

tion

8


DNV GL AS



Sec

tion

8


DNV GL AS





bottom




Sec

tion

8


DNV GL AS


Sec

tion

9


DNV GL AS
























Sec

tion

9


DNV GL AS















App

endi

x A


DNV GL AS


1 General




where








App

endi

x A


DNV GL AS






Point A Point BNo


Ka Kb Ka Kb

1(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250

140 for d le 150150 for

150ltdlt250

160 for d gt 250

128 for d le 150136 for

150ltdle250

145 for d gt 250

160

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

2(1)

128 for d le 150136 for

150ltdlt250

145 for d gt 250


160 for dgt250

114 for dle 150124


127

3

128 134 152 167

4

128 134 134 134

5

128 134 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

6

152 167 134 134

7

152 167 152 167

8

152 167 152 167

9

152 167 128 134

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

10

152 167 152 167

11

128 134 152 167

12

152 167 128 134

13

152 167 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

14

152 167 134 134

15

152 167 152 167

16

152 167 128 134

17

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

18

134 134 134 134

19

134 134 128 134

20

134 134 152 167

21

134 134 152 167

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

22

134 134 134 134

23

134 134 128 134

24

134 134 152 167

25(1)

128 d le 150136 for

150ltdle250

145 d gt 250

140 d le150150 for150ltdle250160 d gt 250

114 d le 150124 for

150ltdle250

134 d gt 250

125 d le150136 for

150ltdle250147d gt 250

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

26

128 134 134 147

27

152 167 134 147

28

152 167 134 147

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

29

134 134 134 147

30

134 134 134 147

31(2)

113 120 113 120

App

endi

x A


DNV GL AS

Point A Point BNo


Ka Kb Ka Kb

32(23)

113 114 113 114





App

endi

x A


DNV GL AS




where




App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


where






App

endi

x A


DNV GL AS




fA = Reduction



= 10 in other cases



App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS




200 9 ndash 13 tw-gr + 245 229 (tw-gr + 09)2

220 9 ndash 13 tw-gr + 276 254 (tw-gr + 10)2

240 10 ndash 14 tw-gr + 303 280 (tw-gr + 11)2

260 10 ndash 14 tw-gr + 330 306 (tw-gr + 13)2

280 10 ndash 14 tw-gr + 354 333 (tw-gr + 14)2

300 11 ndash 16 tw-gr + 384 359 (tw-gr + 15)2

320 11 ndash 16 tw-gr + 410 385 (tw-gr + 16)2

340 12 ndash 17 tw-gr + 433 413 (tw-gr + 17)2

370 13 ndash 19 tw-gr + 475 452 (tw-gr + 19)2

400 14 ndash 19 tw-gr + 517 491 (tw-gr + 21)2

430 15 ndash 21 tw-gr + 558 531 (tw-gr + 23)2


01317hstf- 35726

(tw-gr + 0006hstf- 03139)2




1



080 112

2



100 90

App

endi

x A


DNV GL AS


3



113 80

Quality level C2)


125 72

4



113 80

Quality level C2)


125 72

5



113 80

App

endi

x A


DNV GL AS


6




127 71

Quality level C2)


142 63

7





248 36

Quality level C2)


278 32

8




248 36

App

endi

x A


DNV GL AS


Quality level C2)


278 32

9








10


where



11


Where







90104middotKm3)







App

endi

x A


DNV GL AS


12



Kme = 10Kmα = 10





1


147 61

App

endi

x A


DNV GL AS


2




127

143

71

63

3



18

20

50

45

App

endi

x A


DNV GL AS


4



10

127

18

90

71

50

5


where


147 61

6


180 50

App

endi

x A


DNV GL AS


7




B-C2 150-100

8




10 90

9




09 100

App

endi

x A


DNV GL AS


10



Connection length

Nominal stress

143 63

11




113 80

12


slope 15

slope 13

slope 12


113

127

143

80

71

63

App

endi

x A


DNV GL AS


13


40 23





1



t gt 25 mm and





113

127

119

80

71

76

2



113

127

143

161

80

71

63

56

App

endi

x A


DNV GL AS


3


-Flat bar

-Bulb profile

-Angle profile


160

160

180

56

56

50

4


-flat bar

-bulb profile

-angle profile

160

160

180

56

56

50

5



160

180

200

225

56

50

45

40

App

endi

x A


DNV GL AS


6


Non-load carrying

Load carrying

113

20

80

45




1



90K

App

endi

x A


DNV GL AS


2




90K




1


50 lt d le 100

100 lt d le 150

150 lt d le 300

300 lt d


-reinforced ends or




120

127

133

147

180

75

71

68

61

50

App

endi

x A


DNV GL AS


2


50 lt d le 100

100 lt d le 150

150 lt d le 300


127

132

143

161

71

68

63

56





1


Cruciform joint

Tee-joint

127

113

71

80

2


Cruciform joint

Tee-joint

143

127

63

71

App

endi

x A


DNV GL AS


3



σa=σnmiddott1(2a)

t13 le a

t13 gt a

25

225

36

40

4


t le 25 mm

t gt 25 mm

105

119

86

76

5




App

endi

x A


DNV GL AS









1


Point A

Point B

24

127

375

71

2


Point A

Point B

127

127

71

71

App

endi

x A


DNV GL AS

3


Point A

Point B

117

127

77

71

4


Point A

Point B

117

127

77

71




App

endi

x A


DNV GL AS



1


70 13

2


45 20

3


25 36

App

endi

x A


DNV GL AS




1


For tubular section



161

180

56

50

2


For tubular section



200

225

45

40

3


d le 50 mm

d gt 50 mm


127

143

71

63



App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS




00 90 90 90 90

05 72 80 86 88

10 56 63 75 82

15 50 54 64 73

20 46 50 57 66

App

endi

x A


DNV GL AS




00 90 90 90 90

05 66 72 80 85

10 54 58 65 72

15 49 52 56 62

20 46 48 52 56

App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



App

endi

x A


DNV GL AS


App

endi

x A


DNV GL AS



where



App

endi

x A


DNV GL AS








App

endi

x A


DNV GL AS


App

endi

x B


DNV GL AS



2 Other S-N curves



App

endi

x B


DNV GL AS



App

endi

x B


DNV GL AS




App

endi

x B


DNV GL AS


App

endi

x C


DNV GL AS



where





where






App

endi

x C


DNV GL AS

where








where







cycles

App

endi

x C


DNV GL AS


ξ m = 30 ξ m = 30 ξ m = 30

060061

062

063

064

065

066

067

068

069

070

071

072

073

074

075

076

120000104403

91350

80358

71048

63119

56331

50491

45442

41058

37234

33886

30942

28344

26044

24000

22178

077078

079

080

081

082

083

084

085

086

087

088

089

090

091

092

093

2054819087

17772

16586

15514

14542

13658

12853

12118

11446

10829

10263

9741

9261

8816

8405

8024

094095

096

097

098

099

100

101

102

103

104

105

106

107

108

109

110

76717342

7035

6750

6483

6234

6000

5781

5575

5382

5200

5029

4868

4715

4571

4435

4306


where







App

endi

x C


DNV GL AS

where












App

endi

x C


DNV GL AS




=









060 82 73 66 107113 98104 8995

070 84 76 68 115124 105114 95105

080 85 77 69 118130 108120 98110

090 84 76 69 119133 109122 99113

100 83 75 68 118133 108123 99114

110 81 74 67 117133 107123 98114

120 79 72 66 115132 106123 97113

130 77 71 64 113131 104122 96113

App

endi

x C


DNV GL AS







060 259 233 209 341358 311330 282303

070 227 205 184 309333 282307 256282

080 202 182 164 280308 256285 233262

090 182 164 148 256286 235265 214244

100 165 150 136 236267 217247 198227

110 152 138 125 219250 201232 185214

120 141 128 117 205236 189219 173202

130 132 120 110 193223 178207 163192







060 822 740 662 10831138 9871047 895960

070 612 552 495 831897 758827 690759

080 480 434 391 667734 609677 555623

09 392 355 321 554618 507571 463526

100 331 300 272 472534 433494 396455

110 285 259 235 412470 378435 347402

120 251 229 208 366420 336389 308360

130 225 205 187 329381 303353 279327

App

endi

x C


DNV GL AS


In-Air Δσ0

CorrosiveΔσ0

S-N curves



B1 153 141 130 130 121 112

B 143 132 122 122 113 105

B2 133 123 114 114 106 98

C 118 108 99 99 91 84

C1 106 97 89 89 82 75

C2 94 87 79 79 73 67

D 83 75 68 68 62 56

E 73 67 60 60 55 50

F 65 59 54 54 49 45


App

endi

x C


DNV GL AS



App

endi

x D


DNV GL AS





App

endi

x D


DNV GL AS



where









App

endi

x D


DNV GL AS










where





App

endi

x D


DNV GL AS



where









where





App

endi

x D


DNV GL AS

where







ia ib = IaS Ibℓ



App

endi

x D


DNV GL AS



App

endi

x D


DNV GL AS

where







where





App

endi

x D


DNV GL AS





App

endi

x E


DNV GL AS


1 FE modelling








App

endi

x E


DNV GL AS





App

endi

x E


DNV GL AS













Solid elements


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS

a)

b)


App

endi

x E


DNV GL AS

a)

b)

c)


App

endi

x E


DNV GL AS



where




App

endi

x E


DNV GL AS


App

endi

x E


DNV GL AS













App

endi

x E


DNV GL AS











166

Bulk carrier






App

endi

x E


DNV GL AS




App

endi

x E


DNV GL AS



App

endi

x E


DNV GL AS












App

endi

x E


DNV GL AS







App

endi

x F


DNV GL AS











15 mmt2







App

endi

x F


DNV GL AS



AppA Table 8




tt10 max 15 mm






where




App

endi

x F


DNV GL AS









App

endi

x F


DNV GL AS










App

endi

x G


DNV GL AS




App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS


App

endi

x G


DNV GL AS



App

endi

x G


DNV GL AS



App

endi

x H


DNV GL AS











App

endi

x H


DNV GL AS









App

endi

x H


DNV GL AS


LNG carriers 1000

LPG carriers 1000





















App

endi

x H


DNV GL AS




App

endi

x H


DNV GL AS







bulkheads

















App

endi

x H


DNV GL AS






App

endi

x H


DNV GL AS







LC8 Ballast TBAL

σ LC 8

LC9 Tact σ LC 9

App

endi

x H


DNV GL AS

LC10 Tact σ LC 10

LC11 Tact σ LC 11

LC12 Tact σ LC 12




LC13 Full load Ts

σ LC 13

App

endi

x H


DNV GL AS


LC15 Tact σ LC 15

LC16 Tact σ LC 16





App

endi

x H


DNV GL AS

where





Where





where




App

endi

x H


DNV GL AS

where


= 10 for

= for


= 10 if

= 09 for VLA-E if




Stress range





App

endi

x H


DNV GL AS

where



=




where







App

endi

x H


DNV GL AS



Krsquo in (Nmm2) 6028 6783 6894

n 0117 0111 0115









App

endi

x H


DNV GL AS










where





102 le N lt 104

Materiallog K2 m



App

endi

x H


DNV GL AS



where






App

endi

x H


DNV GL AS


















App

endi

x H


DNV GL AS









Full load-Ballast







App

endi

x H


DNV GL AS


App

endi

x I


DNV GL AS







App

endi

x I


DNV GL AS



Where




Where






Where

App

endi

x I


DNV GL AS






R = route factor








App

endi

x I


DNV GL AS

Where








App

endi

x I


DNV GL AS











App

endi

x I


DNV GL AS


App

endi

x J


DNV GL AS










CONTENTS


Section 1 General

1 Introduction


21 Symbols

22 Abbreviations


3 Application

31 General


33 Vibrations







51 General




55 Stress range


61 General

62 CSA

63 RSD

64 PLUS

65 HMON

66 VIBR and COMF-V

67 NAUT and routing

68 WIV

7 Definitions


1 Introduction

2 S-N curves

21 General







31 General




42 Welded joints

5 Thickness effect

51 General





6 Material factor


1 Introduction


21 General











36 Fatigue life




1 Introduction









33 Stress range

34 Mean stress






51 General





71 General


73 Sign convention









1 General

11 Definition


13 Assumptions





31 Load components






41 Introduction



1 Introduction

11 Thickness

12 Application







21 General

22 FE models

23 Load application

24 Mesh size





31 General








51 General







1 Introduction

11 General

2 Grinding





1 Introduction







1 References


1 General




14 Basis





3 Butt welds





53 Doubling plates




9 Pipe connections





2 Other S-N curves



























1 FE modelling

11 Mesh size

12 Solid elements









51 Introduction

52 Procedure








1 Tolerance limits


3 TIG dressing

4 Hammer peening



1 Introduction




1 General




32 Pressure loads












1 Introduction










11 Introduction


dnvgl-cg-0129 fatigue assessment of ship structures...changes - current class guideline —...

Documents