NTIS #PB95-1 44382 “

SSC-379

IMPROVED SHIP HULLSTRUCTURAL DETAILS RELATIVE

TO FATIGUE

This&cnmmt hasban approvedfcr public release and mlq its

distibutim is unlimited

SHIP STRUCTURE COMMITTEE

1994

SW smuc~The SHIP STRUCTURE COMMllTEE is constitute to prosecutea researchpmgratnto imDrovethe hull structuresof shipsand otfwrmarinestructuresby an extensionof knowledgewfaining to dsslgn,matenalq and methodsof ccmtrucfion.

RADMJ. C. Card, USCG (Chaiian)Chief, Officeof MarineSafely, Security

and EnvironmentalProtectionU. S. Ccast Guard

Mr. Thomas H. Peirce Mr. H. T, Hailer Dr. ConaldIiuhfwimaResearch and Development AssociateAdministratorforShip- SeniorVice President

coordinator buildingand ShipOIMraticmsTranspmtationOevebpment Centw

Am6fican Bureauof ShippingMaritimoAdministratkm

TransportCanada

Mr. EdwardComstock Mr. TnomasW. Allen Mr. Warren NethercoteDirector,Naval Architecture EngineeringOff&r (N7) Head, HydrcfwadbsSectionGroup (SEA OSH) MilitarySealiftCcntmand

Naval Saa SystemsCommandDafence Research Establbhmen-Allantic

CAL RFP&SENI&UvE

COR Steph6mE. Sharp., uSCG Mr. kWliamJ. SiekierkaU. S Ccast Guard Naval sea Syshns Command

me SHIP STRUCTURE SUSCOMMlmEE acb for the Ship SlructuraCommitfw m technicalmattersby provkiingtechnicalcoordinaticmfor determimdingthe oals and objectivesof the programand by evaluatingand inter!xetingthe resultsin termsof

08structuraldesign,camtrmtion. a operation.

MILITARY SFALIFT COMMAND MARITIME ADMINISTRATION ~D

Mr. Robat E. Van JoIws (Chairman) Mr. FrederickSeibdd CAPT G. D. MarshMr. RickardA Anderson Mr. Norman0, Hammer CAPT W. E. Colburn,Jr.Mr. Michael W. Touma Mr. Chao H. Lin Mr. RubinScheinbsfgMr. Jeffrey E. Seach Dr. Walter M. Mac!eaI Mr. H, Paul C@3en

AMFRICAN BUREAU OF SHiPPING NAVALSEAsYSTFMSCOMMANO TRANSPORT CANADA

Mr. StephenG. Amtson Mr. W. Th.wrmsPackard Mr. JohnGrinsteadMr. John F. Cc+km Mr. CharlesL Null Mr. ian BayfyMr. PfdlliPG. Rynn Mr. EdwardKadafa Mr. David L StocksMr. WilliamHanzelek Mr. Ah H. Engte Mr. Peter Imonin

DEFENCE RESEARCH ESTAEi!i NTIC

Dr. Neil PeqLCDR 0.0 #eiliyDr. RogerHofliwsheadMr. Jchn Porter

SHIP STRUCILILRE SUBCO Ml !&SON MEMBERS

u, .s~Y MYOF SCIENCFS

LCDR BruceR. Mustain Dr. Robefl Sbbkl

U. S, MERCHANT MARINE ACADEMY

Dr. C. B. Kim~s

NA NAL @SIFMy SC E

Mr. Peter M. Pafertno

Dr. RamswaI Sh?.ttachaIVya Dr. MartinPrager

C#N,04yCENTR: FOR ~;ERAL S AND

Dr. WMam R. Tysmt

AMERICAN IRON AND STEEL lNSTITLfE

Mr. AlexanderD, WISG+I

S(X EIYoF NAVAL ARCHITECTS ANDM&EWEE5s

Qf@QE OF NAVAL RESEARCH

Dr. WilliamSandberg Dr. Yapa D. S. Ra@paske

~EAL STANDARDS ORGANIZATION

~R

CAPT Charles Piersall Mr. TrevorButlerMemcfid Univwsityof Newfoundland

I1

Member Agencies

Americm Bureau of Shippingknee Research Establishment Atlantic

Maritima AdministrationMiMa Seaiift Command

~Navaf Sea yatema CommandTran.$od Canada

UnifedStatesCoaatGuard

~ c Addraaa Correspmdance !0

Executive DirectorShipStructureComminaa

Ship U.S.coastGuard (G-Ml/SSC)2100 %cond Street, SW.

Structure WSShi ton,D.C.20S%3-0001

CommitteePh:(2032S7-01XX3Fax@02)2S7-4S77

An InteragencyAdvisofyCommittee SSC-379SR-1346

December 22, 1994

IMPROVED SHIP HULL DETAILS RELATIVE TO FATIGUE

As margins for operating ships get tighter and the costs forfailures rise exponentially the need to prevent fractures at thedesign stage becomes increasingly more critical. This reportprovides one more tool for the designer to use. It presents afatigue design methodology that applies existing fatigue data towelded ship details. A variation of the nominal stress approachis used for weld terminations in attached bracket detaile. Thishelps in selecting the weld configurations that improve fatiguelife and assesees the impact of geometric streas concentrationfactors and combined loadings that are typical of welded shipstructural details. Case atudies are shown to demonstrate themethodology. A gloesary of terms used is provided andrecommendations are presented for future research.

Rear Admir$?i, U.S. Coast GuardChairman, Ship Structure’ Committee

.- . .

(THISPAGEINTENTIONALLYLEFT BIANfi

ii

-.. . .

Technicol ReportDocumentationPage

Rep.., No. 2. Gov.rnrns., AccessionNo. 3. Recipient’s Catalog No.

SSC-379 PB95-144382

T,tle.and Subtitle 5, ReportDOI.

September 6, 1994 I

Improved Ship Hull Structural Details Relative to 6, P.rf.rm ina Org.niz.,i.n Cod.

Fat igue

8. P.rlornino O,g.anizationReportN..A.!h.rlx) fQ~l A. Stambaugh, Dr. FredericTkLawrence

and Stimati Dimitriakis SR-1346

‘,Performin9 Org.a. iz.ation Name .nd Addre.s 10. Wo,kUniINo. (T RA!S)

Consulting Naval Architects794 Creek View Road 11. Con,,.., o, Grin? No.

Severna Park, MD 21146 DTCG23-92-C-EO1O28 —13. TYP. o1 Rep.,, o.d Period Covered

2. SponsoringAg.n.y NmmeandAddr.ssShip Structure Committee Final ReportU.S. Coast Guard (G-M)2100 Second Street, SW )4. Sp.nt.,ing Aganey Code

Washington, DC 20593 G-M

5.Supplmen,.ry Noles

Sponsored by the Ship Structure Committee and its member agencies.

6. Ab,t,.ct

This report presents a fatigue design strategy for welded ship structural details.The fatigue design strategy is based on’cumulative damage theory using nominal stres:The approach is modified to account for the complex geometry of welded shipstructural details. Fatigue notch factors and stress concentration factors arederived from experimental data and finite element analysis. Guidance is providedshowing detail designers how to improve the fatigue life of details using thisapproach.

17. Key Words 18. Di.l<ibutien S1at.m..t Distribution unlimited.

Fatigue Available from:Ship Structure National Technical Information ServiceStructural Details U.S. Department of CommerceS-N Curves Springfield, VA 22151

I!

19. Sec.rity Cl.ssif. (of this <.rm.1) 2U. S...rity Clossi(ofof this pq.) 21. No. olPas. s 22. Prit.

Unclassified$27.00 Paper

Unclassified 150 $12.50 Microf

FormDOT F 1700.7 ICI-72) R.p.od.ction of completed POW a.$horizediii.

METRIC CONVERSION CARD

ApproximateConversionstoMetricMeasures

Symbol When You Know Multiply by To Find Symbol

LENGTHioches 2,5 centimeter

:Cn3

feet 30 centimeters cmyd yards 0.9 meters mmi miles 1.6 ki10meter3 km

AREA~2 SJuare inches 6.5 square centimeters cm 2fi2 s@l-tl feet 0,09 squaremeters m 2~dz square yards 0.8 square meters m 2mi 2 square miles 2.6 squsrekilometerskm2

a&s 0.4 heceaes haMASS (weight)

Oz ounces 28 ~S Elb pounds 0.45 ililogl-mns kg

shorttons 0,9 metricion t(21W0 lb)

VOLUME~P teaspcwts 5 milliliters MLT#p tables~ons 15 milliliters mf-

cubkinches 16 milliliters mLfl oz fluid ounces 30 milliliter InLc cm-x 0.24 litem L

pints;t ‘:r- 0.47 liters Lqt quarts 0.95 litem Lgal gallons 3.8 liters Lfi3 cubicfeet 0.03 cubicmeters M3

@3 cubicyards 0.76 cubicmeters m3

TEMPERATURE (exact)‘i= degees subtract32, degrees “C

Fsbsenheitmultiplyby5P Celsius

r“””

ApproximateConversionsfzomMetric Measures

Symbol When YouKnow Multiply by To Find Symbol

LENGTHmm miilimetem 0.04 inches incm centimeters 0,4 inches inm meters 3.3 feet ftm metem 1.1 yardskn3 kilometers

yd0.6 miles mi

AREA~ 2 ~um m~tiekm 0,16 scmsre inches ~2

m 2 @3re meters 1.2 @ale yards yd2

Im3z squm kilometers 0,4 squaremiles mi2ha ht%es 2.5 acks

(10,000m2)MASS (weight)

0.035 Ooncesig i%%&zls

Oz2.2 pmlds lb

t meticton 1J shorttons(1,000kg)

VOLUMEML millilitm 0.03 fluidounces flOzML milliliter 0.06 cubicinches ~3

L liters 2.1 pi33ts ptL liters 1.06qusrt.sL

qtlien 0.26 gallons gal

m3 cubicmeters 35 Cubicfet ft3m3 cubic meters 1.3 cubicyards yd3

TEMPERATURE (exact)“c degees multjjlj::9/5,degm.s “F

Cekius Fahrenheit

40 .20“c

0 20 37 60 80 1001 1 t 1 1

“F I I I I I 1 I I.40 0 32 80 98.6 160 212

water hems lx+ temperature water tmls

——... .

TABLE OF CONTENTSPaae No.

1.0 Introduction, ,, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...1-1

2.0 Fatigue in Ship Structural Details 2-1

2.1 Structural Loading and Stress.. 2-1

2.2 Predicting Fatigue Response .,,,.,, ,, ..,.,..,,,..,.2-3

3.0 Fatigue Design Strategy........,,. ., .,,.,..,.....,.....3-1

3.1 Fatigue Design Stress . . . . . . . . . . . . . ...,...........3-1

3.2 Fatigue Notch Factors, . . . . . . . . . . . . . . . . . . . . . . . . ...3-4

3.2.1 Definition of Fatigue Notch Factors 3-4

3.2.2 Design Curves.........,.,,.,,,. . . . . . . . ...3-9

3.3 Adjustments to Fatigue Life Data 3-11

3.3.1 Mean Stress, . . . . . . . . . . . . . . . . . . . . . . . . ...3-11

3.3.2 Corrosion . . ...3-13

3.3.3 Thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . ...3-13

3.3.4 Fabrication ... .3-14

4.0 improved Details Relative to Fatigue 4-1

4,1 Design Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...4-1

4.1,1 Reducing Fatigue Notch Factors (KJ ., ., 4-15

4.1.2 Reducing Stress Concentration Factors (K~Cf) 4-15

4.1.3 Reducing Nominal Stress 4-15

4.2 Recommended Proportions , . . . . . . . . . . . . . . . . . . . ...4-17

4.3 Application of High Strength Steel 4-23

5,0 Conclusions and Recommendations 5-1

6.0 References . . . . . . . . . . . . . . . . 6-1

Appendix A Analysis of Ship Structural Details Used in Case Studies

Appendix B Development of Fatigue Notch Factors

Appendix C Analytical Prediction of Fatigue Life

Appendix D Glossary of Terms

v.

-—.— —

2-1

3-1

3-2

3-3

3-4

4-1

4-2

4-3

4-4

4-5

4-6

4-7

A-1

A-2

A-3

A-4

A-5

A-6

A-7

A-8

FIGURESPaae No

Typical example of fatigue cracking in shipstructural details.....,.,.. . ...2-2

Definition of fatigue design stress for bracket details 3-2

Estimating fatigue design stress from FEA 3-3

Fatigue design curves developed from ~ 3-10

Ship detail fatigue stress design curve 3-12

Illustration of the relationship between ~ and K,ti 4-14

Frame flange symmetry . ., .,....,. . . . . . . . . . . . . . . . . . ...4-16

Recommended proportions for panel stiffener connections 4-18

Recommended proportions for deep bracket 4-19

Recommended proportions for hatch corners 4-20

Recommended proportions for frame cutout 4-21

Recommended proportions for beam bracket ., 4-22

Double hull tanker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. A-3

Double hull tanker characteristics.. . . . . . . . . . . . . . . . . . . . . . .. A-4

Double hull tanker midship section . . . . . . . . . . . . . . . . . . . . . . ..A-5

Loading on side shell frame cutout . . . . . . . . . . . . . . . . . . . . . . ..A-6

Course mesh FEAmodel ofdouble hull tanker A-7

Detail geometry ofpanel stiffener A-8

FEAmodel ofpanel stiffener . . . A-9

Detail geometry fordeep bracket A-11

vi.

/.&--

-. .

A-9

A-10

A-11

A-12

A-13

A-14

A-15

A-16

A-17

A-1a

B-1

B-2

c-1

c-2

c-3

c-4

c-5

C-6

FEAmodel ofdeep bracket . . . . . . . . . . . . . . . . . . . . . . . . . . .. A-12

Ro/Roside port cutout, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. A-14

Ro/Rosideportcutoutdetail,,. ... . . . . . . . . . . . . . . . . . . . ..A-15

FEAmodel ofside portcutout. . . . . . . . . . . . . . . . . . . . . . . . ..A-17

Double hull barge midship section ,.. .,, . . . . . . . . . . . . ..18 .A-I8

Detail geometry for bulb plate cutout A-19

SWATH ship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. A-2 I

SWATH ship midship section ..,,... .,. ,.. ,., ... A. . . . . .. A-22

Detailed geometry for SWATH ship beam bracket A-23

FEAmodel ofbeam bracket . . . . ,., . . . .. A. . . . . . . . . . . . . ..A-24

FEAofattachment detain . . . . . . . . . . . . . . . . . . . . . . . . . . . ..B-18

FEAofattachment detail il..... ,, .,, ,,, ,,, ., .,, ,,. .. B..B-I9

The predicted effect of bending stresses on the failurelocation for the Primitive F’ - Load Carrying Fillet WeldatlE+07 cycles . . . . . . . ..25 . . . ..C-25

Hardness of the heat affected zone as a function ofbase metal hardness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..C-27

Yield strength as afunctirm of hardness C-28

Comparison of the S-N data for ship structure details and theanalytical expression for the weld Primitive R - Ripple C-29

Comparison of the S-Ndata for ship structure details andthe analytical expression for the weld Primitive G - GrooveWelded Butt Joint...,,.,,,. C-30

Comparison of the S-N data for ship structure details andthe analytical expression for the weld Primitive F -Non-Load Carry ing Fillet Weld.,.,.. ,.. ,.. C. . . . . . . . . . . .. C-3 I

vii.

c-7

C-8

c-9

c-lo

C-11

C-12

C-13

Comparison of the S-N data for ship structure details andthe analytical expression for the weld Primitive F -Non-Load Carry ing Fillet Weld . . . . . . . . . . . . . . . . . . . . . . . . .. C-32

Comparison of the S-N data for ship structure details andthe analytical expression for the weld Primitive F’ -Load Carry ing Fillet Weld . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. C-33

Comparison of the S-N data for ship structure details andthe analytical expression for the weld Primitive T -Weld Termination . . . . . . . . . .,, .,, .,, ., .,, ,,,,.,, ,.C-35

Comparison of the I-P model predictions and S-N data ofship structure details for the weld Primitive F -Non-Load Carrying Fillet Weld, Case 1 C-36

Comparison of the I-P model predictions and S-N data ofship structure details for the weld Primitive F -Non-Load Carrying Fillet Weld, Case 2 C-37

Comparison of the l-P model predictions and S-N dataofship structure details for the weld Primitive F’ -Load Carry ing Fillet Weld . . . . . . . . . . . .,, .,. .,. , . . . . . . . .. C-38

Comparison of the I-P model predictions and S-N data ofshi~ structure details for the weld Primitive T -Weld Termination . . . . . . . . . . .,,,,,,,,,,,,.,,,,,,.,,, C-39

viii.

TABLES

3-1 Basic Weld Configurations and Fatigue Notch Factors . . . 3-5

4-1 Fatigue Notch Factors for Panel Stiffener Connections 4-2

4-1 Fatigue Notch Factors for Beam Bracket ., ., 4-5

4-1 Fatigue Notch Factors for Deep Bracket 4-6

4-1 Fatigue Notch Factors for Flange Transitions. . . . . . . . . 4-7

4-1 Fatigue Notch Factors for Tripping Brackets . . . . . . . . . . . . . . . . . 4-8

4-1 Fatigue Notch Factors for Tee Cutouts . . . . . . . . . . . . . . . . . . . . 4-9

4-1 Fatigue Notch Factors for Angle Cutouts . . . . 4-10

4-1 Fatigue Notch Factors for Bulb Plate Cutouts 4-11

4-1 Fatigue Notch Factors for Deck and Side Penetrations 4-12

4-1 Fatigue Notch Factors for Miscellaneous Cutouts . 4-13

A-1 Stress Concentration Factors for Panel Stiffeners A-1 O

A-2 Stress Concentration Factors of Deep Bracket . . . . . . . . . . . . . . . A-1 3

A-3 Stress Concentration Factors for Side Port Cutout . . A-16

A-4 Stress Concentration Factors for Bulb Plate Cutout . . . A-20

A-5 Stress Concentration Factors for Beam Brackets . . A-25

B-1 Fatigue Strength of Welded Details.. . . . . . . . . . . . . . . . . . . . . . . B-2

B-2 Welded Detail Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . .. B-4

ix.

LIST OF SYMBOLS

a

b

BM

c

D

HAZ

K

&veld

K:ax

\:ax

K

effmax

N

n{

Ni

N,

N,

N,

R

s

SD

Aspp

.

.

.

.

.

.

.

.

.

.

.

.

.

.

Bracket leg length

Fatigue strength exponent

Base Metal

Constant relating to the mean S-N curve

Depth of structural member

Heat affected zone

Fatigue notch factor

Fatigue notch factor for weldment

Value of ~ for axial component of applied stress

Value of K for bending component of applied stress

Value of 1$ representing combined effects of axial and bendingstresses

Geometric Stress Concentration Factor

Elastic stress concentration factor

Inverse slope of mean S-N regression line, also used as exponentcontrolling the thickness effect

Number of cycles corresponding to a particular fatigue strength; totalnumber of nominal stress range cycles also known as fatigue life

Number of stress cycles in stress block i

Number of cycles of failure at a constant stress range

Life devoted to crack initiation and early growth

Life devoted to fatigue crack propagation

Total fatigue life

Ratio of minimum to maximum applied stress

Standard deviation

Log standard deviation of fatigue strength at 10’ cycles

Fatigue strength of a plain plate specimen at a given life (N,)

x.

L

Aswe,d

AS.

s ref

$

$

STa

s“

s,

t

WM

v,

x

a

P

AS~

f)

U’f

u,

‘B

T

Uf

Un

8

AS:eld

AS:eld

.

.

.

.

.

.

.

.

=

.

.

.

.

.

.

.

.

.

.

LIST OF SYMBOLS(continued)

Experimental fatigue strength range of a welded plate specimen at agiven life (N,)

Stress range

Design stress for the reference thickness

Axial component of applied stress

Bending component of applied stress

Applied mean stress

Ultimate strength

Yield strength

Plate thickness

Weld metal

Variation due to uncertaintyof error in stress analysis

in equivalent stress range; includes effects

Ratio of applied bending to applied total stresses

Geometry factor

Number of stress blocks

Design stress range

Limit damage ratio

Fatigue strength coefficient

Local (notch root) residual stress

Bending stress

Shear Stress

Fatigue design stress

Nominal stress

Weld flank angle

Experimental fatigue strength under pure axial loading

Experimental fatigue strength under pure bending loading

xi.

(THISPAGE INTENTIONALLYLEFT BLANK)

xii. L___

-. ,.—

1.0 INTRODUCTION

Cyclic loading causes fatigue cracking in a ship’s welded structural details. If thesedetails are not designed to resist fatigue cracking, the ship’s profitability may beaffected by repair costs and its economic life shortened. Fatigue cracks, for instance,may lead to fractures in ship’s primary hull structure, an event resulting incatastrophic failure. Therefore, designers should use structural details that minimizefatigue damage and ensure structural integrity for the ship’s intended service life.

One technique for predicting and assessing fatigue cracking uses empirical dataderived from laboratory tests of representative structural details. After detailsundergo fatigue tests, test data are analyzed in terms of stress applied to each detailand the number of cycles required to reach failure. The test results are commonlyreferred to as S-N data and are ~resented in S-N curves.

The fatigue design curves presented by Munse (1) and re-analyzed by Stambaughand Lawrence (2) are for various structural geometries that are ditlcult to apply toship structural details. This report presents a fatigue design strategy to apply fatiguedata to welded ship structural details. The fatigue design strategy is based on thenominal stress approach for basic welded structural configurations. A variation of thenominal stress approach is used for weld terminations in attached bracket details.After having separated the global geometric stress concentration factors from thewelded details, it is possible to select weld configurations that improve fatigue life andassess the impact of geometric stress concentration factors and combined loadingstypical of welded ship structural details.

The case studies used to characterize the stress in typical ship structural details arepresented in Appendix A. The approach used to develop the fatigue design strategyis presented in Appendix B, A methodology for evaluating the effect of weldparameters (e.g., geometry and residual stress) is presented in Appendix C, Aglossary of terms is presented in Appendix D.

1-1

“r’

—.—..——

2.0 FATIGUE IN SHIP STRUCTURAL DETAILS

.7..

Throughout its service life, a ship experiences environmental loading which causes cyclicstress variations in structural members. Those variations can cause fatigue cracking inwelded structural details if the details are inadequately designed. A fatigue assessment,supported when appropriate by fatigue analysis, should ensure that structural membersdo not lead to catastrophic failure. Fatigue-critical locations have been identified in asurvey of standard structural details by Jordan et al. in SSC-272 (3) and SSC 294 (4).Stambaugh (5) presents fatigue-critical locations for special details that may lead tofracture. The fatigue life of a structural detail is determined by the number of cyclesrequired to initiate a fatigue crack and propagate it from subcritical to critical size.Description of the fatigue cracking in ships has been documented by Jordan (1) andStambaugh (3). One example of a side shell longitudinal and transverse cutoutconnection is shown in Figure 2-1 (6). This example is one of many that illustrate thecomplexity of fatigue cracking in welded ship structural details. In the example, lateralload from internal cargo and wave impact produces local loads on the side shellIongitudinals. High stress concentrations are produced at the toe of welds in attachedstiffeners and tripping brackets. This, combined with the use of high strength steel,(HTS) produces higher nominal stresses in the longitudinal stiffener (with littlecorresponding increase on fatigue strength) reduces fatigue life to five or ten years atbest. Fatigue analysis should be considered for these locations and wherever special ornew details are introduced in the ship’s primary structure.

2.1 STRUCTURAL LOADING AND STRESS

Hull loads from waves and other sources must be transformed to stress distributions inthe structural detail. Because it depends on the type of ship and operationalenvironment, predicting and analyzing fatigue stresses is complex. The designer mustestimate the magnitude of the stresses and determine their impact on fatigue response.

In a ship’s steel structure, stress cycles are generally caused by the seaway and bydynamic effects such as bottom slamming and hull girder whipping. Changes in cargodistribution and local loads induce bending moments. Together, all of these loadsproduce bending stress and shear stress in the ship’s hull girder. Local stresses causedby changes in hydrostatic pressure and local loading from cargo or ballast are alsosuperimposed on the hull girder stress. If pertinent to a patilcular ship, other loadingfrom dynamic effects, stresses from thermal differences in the girder, and residualstresses should be considered in the fatigue analysis.

Global loads are distributed through plates, girders, and panel stiffeners, all of which areconnected by welded structural details that may concentrate stress.

2-1

.—- ..— .

A LONGITUDINALSTIFFENERCRACKEDB FLAT BAR STIFFENER CRACKEDC SHELL PLATE TO WEB WELD CRACKEDCl CRACK EXTENDINGINTO SHELL PLATED WEB FRAME CRACKEDE BRACKET CRACKEDF LUG CRACKED (TYPICAL DETAIL)

FLAT BARSTIFFENER

BRACKET

WEB FRAMEPLATING

.

,“

. . .

~TYPIC L S1

SIDE SHELL

SIDE SHELLLONGITUDINAL

Figure 2-1 Typical example of fatigue cracking inship structural details

2-2

-

-——-

2.2 PREDICTING FATIGUE RESPONSE

.7.

Loading and resultant stresses are complex and random in nature. Therefore, aprobabilistic approach is often used to characterize the Iong-term stress responsedistribution. The distribution is first developed bycombining probabilities for each loadand corresponding stress state. Then, the stress response transfer function is predictedfor the individual load cases; and, finally, the distribution ofjoint probabilities arecombined based on the probability of occurrence of each sea state.

Techniques for predicting Iong-term load and stress distribution and their developmenthave been investigated extensively by Munse (l), White (7), Wksching (8), and othersbut with little agreement astothe type ofdistfibution that accounts for random loadeffects. The designer, therefore, must choose the dominant loads and combine them asthey areexpected tocombine during theship's semice life. The Iong-term stressdistribution is used inthecumulative damage analysis along with the S-N data applicableto the structural detail in question.

Thecumulative damage approach isamethod used topredict and assess fatigue life.Asdeveloped by Miner (14), this approach requires knowledge of structural Ioading andthestructure’s capacity expressed as stress range and number ofcycles to failure.Developed from test data typically illustrated as (S-N cutves), this method is based onthe hypothesis that fatigue damage accumulates linearly and that damage due to anygiven cycle isindependent of neighboring cycles. Bythishypothesis, thetotal fatigue lifeunder a variety of stress ranges is the weighted sum of the individual Iives at constant S,as given by the S-N curves, with each being weighted according to the fractionalexposure to that level of stress range. To apply this hypothesis, the long-termdistribution ofstress range isreplaced bya stress histogram, consisting ofa convenientnumber of constant amplitude stress range blocks, S1,and a number of stress cycles, n,.Theconstraint against fatigue fracture is then expressed interms ofa nondimensionaldamage ratio, rF

;=l IV(

where B=n, =Ni =Ll =

The limit damage ratio~,

number of stress blocksnumber of stress cycles in stress block inumber of cycles of failure at a constant stress range. S,Iimit damage ratio

denendson maintainabilitv. that is, the Dossibilitv forinsDectionand repair, and the fatigue characteristics of the particular detail. These factors alsohave probabilistic uncertainty associated with them.

2-3

Fatigue design, using the Iinearcumulative damage approach, ensures the safety orpeflormance ofasystem foragiven period oftimeand/or under a ''specified'' loadingcondition. But the absolute safety of the system cannot be guaranteed because of thenumber of uncertainties involved. In structural design, these uncertainties can be due tothe random nature of loads, simplifying assumptions in the strength analysis, materialproperties, etc.

2-4

3.0 FATIGUE DESIGN STRATEGY

A fatigue design strategy is presented to facilitate correlation between existing fatiguedata and welded ship structural details. The fatigue design strategy is based on fatiguedata presented by Munse (1) and re-analyzed by Stambaugh and Lawrence (2) forvarious structural geometries. Fatigue response data are presented to use withgeometric stress concentration factors and combined loadings typical of ship structuraldetails as developed in Appendix A and B. The fatigue design strategy is based on thenominal stress approach with modifications for induced stress concentration factors (e.g.,brackets, toes and weld terminations) with various geometries. After having separatedthe global geometric stress concentration factors from the welded details, it is possible toselect weld configurations that improve fatigue life and assess the impact of geometricstress concentration factors. A methodology for evaluating the effect of weld parameters(e.g., geometry and residual stress) is presented in Appendix C.

3.1 FATIGUE DESIGN STRESS

Fatigue design stress (~,) is defined as the stress range (double amplitude) in theIocationof the weld intheabsence of the weld. Theoverall geometry of the weld neednot be considered unless there are discontinuities from overfill, undercutting, or grossvariations in the weld geometry. The relevant stress range must include any localbending and stress concentrations caused by the geometry of the detail as describednext.

For bracketed details, combined stress from vatious load sources (shown in Figure 3-l)can be obtained from Finite Element Analysis (FEA). The maximum principal stress(9)should be used for combined stress fields. For deep beams and girders, bending stressis essentially an axial stress at the location of interest. This is in contrast to platebending andassociated gradients that have aneffect onthe fatigue life. Where out ofplane stresses are high, the maximum principal stress may occur at the upper weld toein the attachment. Thus, knowing where the maximum principal stress occurs isimportant and can be identified from FEA.

An illustration ofglobal geometV andlocal weld toegeomet~ isshown in Figure 3-l.Stress associated with thephysical geomet~in structural details can reestimated byFEA. Thestress gradients areveVsteep inthevicinity of the weld toe. Because of thehigh gradients, the maximum stress computed or measured will be sensitive to the meshsize. Because ofthis mesh sensitivity the fatigue design stress developed using FEAmust be defined. The fatigue design stress is the principal stress on the order of oneplate thicknesses from theweld toe as illustrated in Figure 3-2. Parametricapproximations of stress concentration factors can be used to screen details; however,FEAshould beusedfor fatigue critical locations. Theapplication of the finite elementtechnique to ship structural details is described by Liu and Bakker (10).

3-1

-“

b ‘Jf

+

AL-t

+

~f=U. . FROM FEAprmctpal

C7f = ~ ● K=cfn FROM BEAM ANALYSIS

Figure 3-1 Definition of fatigue design stress for bracket details

3-2

L--

--

PEAK STRESS

II

I

FATIGUE DESIGN SIRESS

Of t AWAY FROM WELD TOE

IIIIII

It WELD TOE (NoT MODELED)

u l+-i--

WELD TOE LOCATION

*t

Figure 3-2 Estimating fatigue design stress from FEA

3-3

3.2 FATIGUE NOTCH FACTORS

Fatigue Notch Factors (K,) associated with basic weld details provide a valuable tool inassessing the fatigue life of welded ship structural details because they can be used inquantitative evaluations and comparisons. Clearly, this is beneficial for application tovarious geometries of welded ship structural details, Baseline fatigue notch factors aredeveloped that represent butt welds or fillet welds. Inthiscase theeffect of the localstress concentration at the weld toe is included in ~, Therefore, the fatigue notch factorincludes effects associated with weld geometry.

3.2.1 Definition of Faticme Notch Factors

The basic weld configurations presented in Table 3-1 are correlated to a basic shipstructural detail design curve using a fatigue notch factor Kf.

The fatigue notch factor Kt for each detail was estimated from the University of IllinoisUrbana-Champagne (UIUC) fatigue data bank (2),(11) information in the followingmanner. At a given fatigue life, the fatigue notch factor ~ is defined as:

~, . il.%mooth specimenf (2)

ASwe/dmenf

The ratio of mean fatigue strength at 10e cycles of smooth specimen to that of plain plateis 1.43. Therefore, the K, can be written as:

~ . 1,43 ASp/ain platef ASwe/dment

~ ~ 1,43 ASp/ain p/atef at 106 cyc/esand for R=O

A.%eldmenf

The development of fatigue notch factors is presented in Appendix B.

(3)

(4)

3-4

_–—-..– x-

b=II

b-

.. c

Table 3-1Basic Weld Configurations and F@igue Notch Factors (con’t.)

U, = an

Weld Detail Description Fatigue Design StressafAxial

-mP- ‘Syrse’y’oadedbu” 246 “e:

Transversely loaded2.63 2.63

groove weld.

& ‘:;=’y’oaded’”e’ 252 293

u’ = u“

U, = an

l“”.,—

Table 3-1Basic Weld Configurations and Fatigue Notch Factors (con’t.)

Weld Detail Description

-fp

1K

@-

Longitudinally loaded weld

I termination.

L- l’J

/ N

Axial and lateral (out of

plane) loaded fillet weld

I

Axial

3.6

3.0

5.5

Bending

3.6

3.0

4.4

Fatigue Design Stress Uf

Right angle connection

using nominal stress (Un) in

base member and no load

in attachment. Axial and

bending are the same for

attachments to deep

sections.

Stress at one t from weld

toe with variable geomet~

and combined stress from

reaction in attachment (at).

Jse this L when out of

)Iane axial and bending

;tress are much greater

han in plane stress. Root

‘ailure is likely for axial load.

T“

Table 3-1Basic Weld Configurations and Fatigue Notch Factors (con’t.)

L===

.,r.

Description

Lap weld in plane load.

Lap weld out of plane

load.

Axial

2.91

5.5

Bending

2.91

4.4

Fatigue Design Stress q

~f = nominal stress at t

away from weld toe. Use &

for axial load in bending.

Axial load induces bending.

Uf = nominal stress at t

away from toe of weld.

Axial load induces bending.

I

_.. .

3.2.2 Desicm Curves

The mean fatigue strength of a weldment based on its fatigue notch factor and thefatigue strength of the plan plate specimen at the fatigue life in question can be writtenas:

AsAswe,d = 2 =

1.43 Asp,

K fwe[d K hveld

(5)

Assuming that the scatter in fatigue life data can be described by the standard deviationof the log of the fatigue strength (SD), the design stress would be:

Asde,i,n = Asw,d -2 ‘f cl‘“

where:SD = Log standard deviation of fatigue strength at 106 cycles

Thus, at 10s cycles

Asde*,gn =‘1.43 As,p

240SDK-hveld

(6)

(7)

This relationship is illustrated in Figure 3-3.

The curves are assumed to be parallel consistent with recent work (2) and currentpractice in development of fatigue design curves (12, 13) for welded structural details.

The approach used to develop the 1$curves and data is discussed in Appendix A. Thewelded detail 1$description, loading, and pictographs are presented in Table 3-1.

The basic design curves, which consist of linear relationships between log (ASR) and log(N), are based on a statistical analysis of experimental data as described by Stambaugh(2). Thus the basic design curves are of the form:

log (N) = log C - m log (AS.)

3-9

I I

(/74

O-JInalL

z

I

I I Weldment (Mean–2SD)

Nt=l ,000,000 cycles

Total Fatigue Life, Nt(cycles)I

Figure 3-3 Fatigue design curves developed from 1$

[

,.

~—.--

or in terms of stress range:

AS, = (C/N) “mwhere:

N is the predicted number of cycles for failure under stress range AS~C is a constant relating to the mean design curvem is the inverse slope of the design curve

The fatigue design curve shown in Figure 3-4 includes the mean minus two standarddeviation adjustment. The relevant statistics are:

Iogc = 4.38m =3

SD = .0696 at N 10’ cycles

The slope of the design curve is hi-linear to account for the constant amplitude fatiguelimit. This limit begins at 5.108 cycles. When all nominal stress ranges are less than theconstant amplitude fatigue limit for the patiicular detail, no fatigue assessment isrequired.

The design curve has a cut off limit at 108 cycles. This limit is calculated by assuming aslope corresponding to m=5 below the constant amplitude fatigue limit. All stress cyclesin the design spectrum below the cut off limit may be ignored when the structure isadequately protected against corrosion.

Other than as described above, no qualitative adjustments are included in this data set.Adjustments required to account for other factors influencing fatigue response are Iefl tothe designer, who should find the research described in the following sections helpful.

3.3 ADJUSTMENTS TO FATIGUE LIFE DATA

3.3.1 Mean Stress

The correction for mean stress ratios other than R=O is based on work by Lawrence(13), who propose an equation to calculate the mean fatigue strength of weldments atlong lives.

AsR 1+(2t$b

AsR=o - ,.1 +R (8)--(ml’

3-11

-.—. —

/’

L--n

II

m0 -0

(Ddll) ‘SV %3NWI SS2UlS

3-12

m0

ho

In

o0

mo

+“o

00-

c0.—In

%

.(u

IiiuQ..—cCn

-— -.———

This equation is used to predict the mean fatigue strength at any R value at 10e cyclesfrom the R=O fatigue strength at 106 cycles, Fatigue strength exponent b is estimatedby:

50b=-:/og2(l+—1.5s”)

(9)

where SUis the ultimate strength of base metal. The derivation of this correction ispresented by Stambaugh and Lawrence (2) along with its validation using the UIUCfatigue data bank.

3.3.2 Corrosion

Salt water can seriously affect the fatigue life of structural details, The data available(15), (16), (17) indicate that corrosion decreases fatigue life where details are uncoatedor do not have cathodic protection. When no consistent protection is provided, evidencesuggests that fatigue life should be reduced by a factor of two for all categories.Corrosion also affects fatigue limit, which becomes non-existent when corrosion ispresent. As noted by UK DOE (18), the design curve must be continued without achange in slope.

3.3.3 Thickness

At present, most agree that for geometrically similar welds larger weldments will sustainshorter fatigue lives. Theoretical (19) and experimental (20) evidence confirm theexistence of a size effect, but there is much scatter in the data. Thus, the magnitude ofthe thickness effect remains in question. Lawrence (11), Gurney (21), and Smith (22)recommend the following relationship:

[1[1s, t, mq’~

(lo)

where t, is taken to be 25mm (1 inch)t, is the thickness of plate (mm)S, is the design stress at the thickness in questionS, is the design stress for the referenced thicknessm is 1/4 as recommended by Lawrence (11) for the S-N curves given

by Stambaugh and Lawrence (2).

3-13

_._. .—. -.

The 25mm reference thickness cited is greater than most structural details constructed ofsteel plate and shapes. Therefore, the correction need not be applied unless the baseplate thickness is greater than 25mm.

3.3.4 Fabrication

The fabrication process is a very important factor in the fatigue life of welded structuraldetails. Data used to develop the fatigue design strategy assume that weld quality isfree of critical defects and meets the requirements of regulatory and classificationsocieties. Joint misalignment has a significant effect on fatigue life (23),(24). Weldprotile changes by grinding and peening affect fatigue response as noted in the UK DOE(18) design code. Residual stress is a very important factor especially in weldtermination. Control of weld geometry and residual stress are effective means ofincreasing fatigue life. The analytical expressions presented in Appendix C can be usedto assess the impact of weld parameter control on fatigue response. Although weldparameter control is often considered expensive, it is worth considering in special cases.

3-14

4.0 IMPROVED DETAILS RELATIVE TO FATIGUE

..v ..–

Ship structural detail design depends on many factors that are unique to the specific

application. Ship tYPe, size, loading, detail location and many other variablesinfluence their design. However, basic parameters can guide detail designers inselection and application of structural details. These parameters include weldconfiguration, detail geometry and nominal stress. An understanding of theseparameters and their relationship will aid in selecting, evaluating and finalizing detaildesign as described next.

4.1 DESIGN OBJECTIVE

The approach based on 1$ can be used by designers to improve fatigue life of weldedship structural details. Separating geometric effects (KC,) from the fatigue notchfactor (KJ enables ship structural designers to control variables that influenced fatigueresponse. The designer can determine which parameters he must control within hisdesign constraints (cost and construction capability) when the primary objective is aconstant fatigue life for a specific detail. To illustrate this point, the fatigue life (N)based on & and K~C,can be expressed as:

N = f (u,,KJ

where; at = crn for simple geometries and

a, = m, * K,Ct for more complex geometries (e.g. brackets)

here; an is the nominal stress and

m, is the fati9Ue design stress one plate thickness from the weld toe

Assuming the designer is working to a constant fatigue life, the important parametersbecome ~, K,ti, and cm. As a practical matter, it is very difficult to design shipstructures using K*C,because it varies depending on application and FEA is requiredto determine the fatigue design stress c, for fatigue critical locations. All too oftendetail designers are expected to provide a detail (K~d) that will improve fatigue life;however, K,Cfalone is insufficient and re-evaluation of the nominal stress an isrequired in many instances. Nominal stress has a significant influence on fatigue life.Detail designers must assess the trade-off between these parameters because theselection of details depends on the specific application. The reliability approachdeveloped by Munse (1) and t$ presented in Table 4-1 provide guidance in makingthis assessment when combined to illustrate the trade-off between 1$ and K,Cr Thefollowing can be inferred by inspection of the information provided in Figure 4-1.

4-1

Table 4-1

Fatigue Notch Factors forPanel Stiffener Connections

Ship Detail K, Comments

&

Connection has high stressconcentration factor and is

3.0 suitable for low nominal

+0

stress applications. K~ti of

+ 3,3 or greater,

&

Connection increases area3.0 and reduces stress concen-

+0

tration slightly. ~d of 2.8.

+

$!&

Connection area and bracketreduce stress at bracket toe.

3,0K~Cfof 2.7, Fatigue criticallocation depends on effective

+0 +shear connection tolongitudinal.

A

K,Cfof 2.3. Fatigue critical

3.0location depends on effectiveshear connection to

+0 +longitudinal.

4.2

L-

Table 4-1(Cont.)

Fatigue Notch Factors forPanel Stiffener Connections

Shin Ilatail I K. Comments-.. .r ------ --r

4!!s Straight brackets reduce3.0 overall stress in connection.

However, K5&of2.7ishi9h.

+0 +

A

Double radius bracket is

3.0required when using HTS.Seediscussion in report. tQ

+0 +

of 2.0.

F

Shear connection between

2.46longitudinal and transversemust be evaluated for

+y\+specific cutout.

Ir+jlOut of plane bending on fillet

4.4welded attachment increases&significantly. &fandanshould be evaluated carefully.

4.3

_ ..._

Table 4-1(Cont.)

Fatigue Notch Factors forPanel Stiffener Connections

Ship Detail K Comments

&

Lapped attachments haveslightly higher 1$ than landed

2.62attachments. Thisconnection introduces high

+ 1- +K,cr Use for low stress (mn)applications,

&

Fatigue critical location2.62 dependson effective shear

+1 +

connection to longitudinal,

Asymmetrical flange

I[%ji

introduces out of planebending from shear center

4.4 load center offset,Corresponding ~ is highreducing fatigue life. Use inlow stress (an) applications.

4.4L-

-.v

Table 4-1

Fatigue Notch Factors forBeam Bracket

Ship Detail K Comments

v

+

Lap brackets generally have

2.91higher out of plane inducedloading, Snipe flange toreduce K~Cfat flange end.

+

T

+

2.91Radius bracket reduces &&See Figure (4-6) for details.

Y

v

+

Flanged brackets have higher

3.0K@ than plain but are moresusceptible to buckling if notdesigned correctly.

Y

F

+

Radius reduces K.Cti Shape

3.0flange 5:1 slope to reduceK~cP See Figure (4-6) fordetails.

Y

4.5

Table 4-1

. _.

Fatigue Notch Factors forDeep Bracket

Ship ”Detail K Comments

b

Stiffener at end of bracket

3.0introduces high & UseFEA for high stress

+ +

applications.

L

Most economical means ofreducing && See Figure 4-3

3.0 for recommended propor-tions. Use FEA for high

+ +stress applications.

L

Slight increase in K,ti Use3.0 FEA for high stress

applications.

+ +

L

Best configuration to reduceK,d at bracket toe. Also

3,0 reduces stress from out ofplane bending at toe. Exact

+ +“

geometry should bedetermined using FEA.

4-6 -

_.._.

Table 4-1

Fatigue Notch Factors forFlange Transitions

.Chin Ik4ail K. I Comments-,..~ -“..... --l

~

) Tapered flange slope must

J

2.58be >5:1. Difference in

flange widths should be* evaluated carefully.

F4

2.04Weld quality is important tomaintain low &

%

4.7

T-

Table 4-1

Fatigue Notch Factors forTripping Brackets

Ship Detail K Comments

A

1

Straight bracket has high K,CP

Kd = 2.7. Effective shear3.0 connection between

longitudinal and transverse is

-+1 -+ very important.

A

This configuration reduces3.0 ~Gf at bracket toe; however,

heel has high &.

-+0 -)-

&

3.0Heel bracket reduces K=,slightly.

-- (-u -)-

?!!!&

3.0 K=, = 2.0.

-+0 +

4.8L-

Table 4-1

Fatigue Notch Factors forTee Cutouts

Ship Detail K Comments

m

1

No effective shear connection1 -1.7 is provided on the open

cutout. This increases U, at

2point 1 and 2. Should be

2- 3.0 considered for low stressapplications.

m

11 -1.7

2It is important that the lug

2-2.62connection be designed to

3transfer shear withoutincreasing Uf at point 2.

3-3.44

m

1

1 -1.7 Most effective method oftransferring shear to thetransverse structure. This

2 2-3.44 reduces q at point 1.

m

1

1 -1.7 Note increase in attachmentlength at web reduces K,Cfatpoint 2 and shear stress

2 2-3.44 across the attachment.

4.9

——

Table 4-1

Fatigue Notch Factors forAngle Cutouts

Ship Detail K Comments

d

1

2Kti (Ref. 23)

1 -1,7 1-2.19

2 -1.7 2 -4.5

M

2

1 KC, (Ref. 23)

1 -1,7 1 -4.4

32 -1.7 2 -3.3

3 -3.0 3-49

M

2

1& (Ref. 23)

1 -1.7 1 -3.7

3 2 -1.7 2 -2.8

3 -3,0 3 -4.1

1 -1.7 & (Ref. 23)

2 -1.7 1 -3.5

3 -3.0 2 -2.4

4-2.62 3 -4.0

4-1o-

_.–. —

Table 4-1

Fatigue Notch Factors forBulb Plate Cutouts

._

.m,,~ “=. -,. .-!

m Small radius increases &&

1.7Note lack of shear transfer totransverse. Use in low stressapplications.

m

11 -1.7

Geomety must be evaluated

2carefully to reduce K~cP

2-3.44

m1 1 -1.73 Effective shear connection is

2-3.44 important in reducing nominal2 stress at point 3.

.,3-2.62

~

Weld wrap and quality of2.93 weld are important in tight

connection.

CL.:.. ila+ail K. Commente

II

4.11

Table 4-1

Fatigue Notch Factors forDeck and Side Penetrations

Ship Detail & Comments

T

1,7

m

Face plate introduces weld

3.0increasing 1$ but reduces IQin detail. Weld quality is veryimportant in this area.

A

KC, is very sensitive to

Bj 1.7opening size and radius,

A: See refs. (25) and (26) forexamples.

El

a

Weld quality is very important

3.0 for all main deck and bottompenetrations andattachments.

4-12L—.

Table 4-1

Fatigue Notch Factors forMiscellaneous Cutouts

Ship Detail K, Comments

m

Size and number of cutouts

3.0are important relative toadjacent structure and canincrease frf at critical location.

a

3.0

m

3.0

EZJ

1.7

4.13

—.. . ...-_—.

4- -

3- -

Kf

2- -an<.3uy HS

— — —— —— __ _ – Kf=l.43 PP

1- -

1 2 3 4

‘SCF

N=l 08 CYCLES

Figure 4-I illustration of the relationship be~e.en ~ and K=,

4-74

K,,, * & <4 for High Strength Steel (an = .3uY HS)

K=, * 1$ <8 for Ordinary Strength Steel (an = .3uY OS)

While these are approximate relationships, they are useful in comparing details andevaluating the trade-off between K,Cf, Kf, and an. Final determination of Uf should bebased on FEA and & presented in Table 3-1.

4.1.1 Reducinq Fatique Notch Factors (Kf)

Improvements in 1$ result from changes in weld type, weld geometry, residual stressor mechanical profiling. The effects of these parameters can be significant and usedas a technique to improve fatigue life. Weld profiling by grinding and peeningimproves Kf and extends fatigue life. These techniques are generally used selectivelybecause of there associated increase in fabrication cost, Analytical expressionsinvolving these parameters and effects on Kf are discussed in greater detail inAppendix C. Typical values of K, are presented in Table 4-1 for ship structural detailsbased on inspection of the details and application of ~ values from Table 3-1.

4.1.2 Reducinq Stress Concentration Factors (KSCf)

Stress Concentration Factors (K,C,) have an infinite number of variations. Thedesigner can select from a number of geometries each of them having a significanteffect on the fatigue design stress Ur Table 4-1 presents typical values of K,d toillustrate the trade-off between ~ and K.=t. The & KWfcurves shown in Figure 4-1can be used to screen details and aid the detail designer. Final selection of the detailshould be based on FEA to determine cre

4.1.3 Reducinq Nominal Stress

Reducing nominal stress in ship structural details is an effective way to reduce fatiguedesign stress (u,) and improve fatigue life.

For example, an increase in frame section modulus will reduce the stress in the detailand weld toe, assuming constant load (which might be typical in using design rules).Similarly, reduction in stiffener or frame span and spacing will reduce nominal stress.The nominal stress in the structure has a significant influence on the fatigue designstress (af) and fatigue response. Therefore, fatigue evaluations should be conductedearly in the ship design because structural detail geometry produces stressconcentrations that cannot compensate for detrimental effects of high nominal stress.

Another example is shown in Figure 4-2 for the symmetry of the flange onIongitudinales. The flange symmetry has significant influence on fatigue strength. Itwas reported that a second generation VLCC experienced fatigue cracks in

4-15

_,... ,,.——.—

T I

M FLAT BAR LAPPED ONANGLE, CUT CHANNEL ORBULB ANGLE

EdFLAT BAR LAPPED ONBUILT UP ANGLE

FLAT BAR BUTT TOTEE

TYPICAL WEB FRAME PANELSTIFFENER CONNECTION TO SIDELONGITUDINAL

Figure 4-2 Frame flange symmetry

4-16

asymmetricflanges after three to four years (27). There were no fatigue cracksfound in a similar ship with symmetric flanges. An investigation found that themaximum stress in the asymmetric configuration is nearly 70 percent higher than insymmetric flanges. Therefore, use of symmetric Tee sections reduces a componentof nominal stress and improves fatigue life.

4.2 RECOMMENDED PROPORTIONS

Numerous examples are provided in Table 4-1 showing the trade-off between K,CfandK for panel stiffeners, tripping bracket connections, frame cutouts and for shellcutouts. Structural detail proportions are very important in lowering K,Cfand K.Recommended proportions are shown in Figures 4-3 through 4-7 based on theanalysis presented in Appendix A.

Recommended panel stiffener ends proportions are presented in Figure 4-3. Bothtoe and heel brackets are required to achieve a K~Cfof less than 2.0.

Recommended deep brackets proportions used in double hull tankers are shown inFigure 4-4. The extended bracket toe radius reduces out of plane stress at the weldtoe.

Recommended hatch corners and side shell cutouts proportions are shown in Figure4-5. The exact proportions of these details depend on the specific application(25),(26).

Recommended bulb plate stiffener cutout proportions are shown in Figure 4-6. Thereare a large number of variations in cutout geometries and Table 4-1 shows K,cf forvarious angle cutouts based on data for standard structural arrangements (24).Additional proportions for cutouts are provided in SSC-266 (26). Generally, smallradius corners should be avoided. Effective shear connections are extremelyimportant in reducing K,=f in cutouts.

Recommended beam bracket proportions are shown in Figure 4-7. A commonfeature seen in the figures described above includes 5:1 slope on shaped flanges toreduce K,& Generally, plain brackets have lower &c, than flanged brackets; however,plain brackets are more susceptible to buckling. Straight brackets are shownbecause they are more common than radiused brackets. Radiused brackets havemuch lower K,ti than straight brackets and are worth considering for plain brackets

4-17

--?-

—d—

“ - t-

(1.5t flange ma.)

t

2/3d

(30; rein)

L-

min 12 (tankers)

~ L (1.5t flange max)

KSCFs 1.35

Figure 4-3 Recommended proportions for paneI stiffener connections

4-18

A

a

/ VA VA

k5

slope > 5.1

\\

min

(y’ t flange

L ~ 5t rnin

t~ 1.5t max

1

t)t

&F ❑ 1.52

Figure 4-4 Recommended proportions for deep bracket

4-19

t

R >300KS= VARRIES DEPENOINGON APPLICATION. SEEREF 25 ANO 26.

$

B> ’2A

300(MIN) A

& x 1/2 RADIUS CORNERFOR SIMILAR APPLICAllONS.

K~w N SAME AS ELLIPSE.

1I

J

HATCH CORNERS

Figure 4-5 Recommended proportions for hatch corners

4-20 L___

_–— ....

A A Av VA v v

“’---

+ b/2 min

iJ I

BULB FLAT

Figure 4-6 Recommended proportions for frame cutout

4-21

——

b“ ‘

i dd

—“---i

r J

L 1.5t BRACKET MAX

a

Ln‘d

a>l .5d/’ 1

/’

P

MIN. SLOPE

/’ 5/’

/’

K,c, VARRIES DEPENDING ON. .DEPTH OF BRACKET ANOSUPPORTiNG MEMBERS, FLANGEOBRACKETS HAVE HIGHER KWTHAN UNFLANGED.

Figure 4-7 Recommended proportions for beam brackets

4-22

not requiring rolled flanges. Propoti[ons for panel stiffener connections and deepbracket may be used for radiused brackets. Recommended proportions for bracketthickness, leg length and flange size is presented by Glasfeld (26) and the TankerForum (28).

It is extremely important to use good fabrication practices described by Jordan (25)when using the fatigue design strategy and recommended proportions presented inthis report. The depth of bracket ends (t~l .5) is extremely important in maintaining aK of 3.0.

Clearly, there are various improvements that reduce K,Ct The final selection ofdetails and determination of ~f must be verified by the designer using FEA for specificapplications. The cost trade off must be assessed by the designer based on savingsof material, labor, and shipyard resources. A guide for estimating the cost ofstructural details is provided by Jordan in SSC-331 (29).

4.3 APPLICATION OF HIGH STRENGTH STEEL

The application of High Strength Steel (HTS) in ships must be approached carefully.Although the yield strength of HTS is greater, the fatigue strength of welded structuraldetails is approximately the same as ordinary strength steel. When scantlings andresulting section modulus are reduced the nominal stress increases. This translatesto an increase in nominal stress at the connecting details. This must becompensated by using details with reduced K,ti For example, in sizing side shelllongitudinal stiffeners of AH-36, the section modulus can be reduced to i’z~. ofordinary strength steel based on the high strength steel factor, Q=.72, by ABS (28).This produces a 40% increase in stress at the detail (assuming constant load). Thegeometric K~d must reduce the stress by 40% to maintain constant fatigue life. Byinspecting the trends in K,ti shown in Table 4-1, the double radius bracket is the onlydetail that produces more than 40% reduction in K,ti over straight panel stiffeners.The designer may also choose a smaller increase in nominal stress (say 20Yo) andcompensate with a detail that reduces the K~&by 20V0. This trade-off depends oncost for the specific application, Figure 4-1 illustrates the trade-off between K andK~Cffor ordinary and high strength steel. If t$ and K,d are to the left of the respectivematerial curve, the detail is satisfactory for the nominal stress indicated. If not, thenominal stress should be re-evaluated or detail & or K,,f changed.

4-23

‘1””

.

5.0 CONCLUSIONS AND RECOMMENDATIONS

1. Recent advances in computer technology and development of pre-processors for finite element programs allows designers to analyze thestresses in ship structural details quickly. Variations can be evaluatedand parametric analysis of detail configurations can be performed toguide the designer in assessing fatigue critical details. However, similartechniques are required to guide the designer in developing loadhistories quickly. The reliability approach developed by Munse (1) canbe applied easily; however, its application has not been verified andcalibrated for general use. Further development of this type ofapproach, combined with the fatigue design strategy presented here, willexpedite detail design and fatigue analysis of more details requiringattention by designers.

2. The fatigue design strategy presented here should be used to re-evaluate stiffened panel design criteria in Iightof the fatigue notchfactors and stress concentration factors for typical welded structuraldetails. This evaluation should include the effects of high strength steeland non-linear effects of torsion in panel stiffeners.

3. The approach used to predict effects of weld parameters for weldterminations has been developed using existing data for attachments;however, the technique should be verified for combined loading andsheer loading typical of terminations found in welded ship structuraldetails. This effort should include both testing and analythd evaluations(using FEA) of the test specimens. Three dimensional effects at theweld should be evaluated both experimentally and analytically.

5-1

---

(THIS PAGE INTENTIONALLY LEFT BL4NK)

~-”

6.0 REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

Munse, W. H., Wilbur, T.W., Tellalian, M. L., Nicolle, K., and Wilson, K.,“Fatigue Characterization of Fabricated Ship Details for Design,” SSC-318, 1983.

Stambaugh, K., Lesson, D., Lawrence, R., and Banas, “Reduction of S-N Curves for Ship Structures,” SSC-369, 1992.

Jordan, CR. and Cochran, C. S., “In-Service Performance of StructuralDetails,” SSC-272, 1978,

Jordan, C.R. and Knight, L.T., “Further Survey of In-ServicePerformance of Structural Details,” SSC-294, 1980.

Stambaugh, K. and Wood, W., “Ship Fracture Mechanisms investiga-tion,” SSC-337, March 1987.

Exxon Corporation, “Large Oil Tanker Structural Survey Experience,”Position Paper, June 1, 1982,

White, G.J. and B.M. Ayyub, “Reliability Based Fatigue Design for ShipStructures,” ASNE Journal, May 1985.

Wirsching P. H., Chen Y.-N., “Considerations of Probability-BasedFatigue Design for Ship Structures,” ASNE Journal, May 1985.

Stambaugh, K. and Munse, W. H., “Fatigue Performance under Multi-axial Loading Conditions,” SSC-367, 1990.

Liu, D. and A. Bakker, “Practical Procedures for Technical andEconomic Investigations of Ship Structural Details,” Marine Technology,January 1981.

Lawrence, F.W., “Fatigue Characterization of Fabricated Ship Details --Phase Ii,” Ship Structure Committee Project SR-1298, University ofIllinois, Urbana, Illinois (awaiting publication).

“Guidance for the Survey and Construction of Steel Ships,” Nippon KaijiKyokai, 1989.

“Recommendation for the Fatigue Design of Steel Structures,” ECCS,1985.

6-1

14. Miner, M.A., “Cumulative Damage in Fatigue,” Journal of AppliedMechanics, Vol. 12, 1945.

15. Marshall, P., “Basic Considerations for Tubular Joint Design in OffshoreConstruction,” Welding Research Council Bulletin 193, April 1974.

16. Burnside, O.H., S.J. Hudak, Jr., E. Oelkers, K. Chen, and Dexter R.J.,“Long-Term Corrosion Fatigue of Welded Marine Steels, “SSC-326,1984,

17. Albrecht, P., Sidani M., “Fatigue Strength of Weathering Steel forBridges,” University of Maryland Department of Civil Engineering,October 1987.

18. U.K. Department of Energy (DEn), “Offshore Installations: Guidance onDesign and Construction,” January,l 990.

19, Gurney, T. R., “The Influence of Thickness on the Fatigue Strength ofWelded Joints,’r Proceedings 2nd International Conference on Behaviourof Offshore Structures (BOSS), London, 1979.

20. Maddox, S.J., “The Effect of Plate Thickness on the Fatigue Strength ofFillet Welded Joints,” The Welding Institute, 1987.

21. Gurney, T. R., “Revised Fatigue Design Rules,” Metal Construction 15,1983.

22. Smith, I.J,, “The Effect of Geometry Change Upon the Predicted FatigueStrength of Welded Joints,” Proc. 3rd Int. Conf. on Numerical Methodsin Fract. Mech., pp. 561-574.

23. General Dynamics Corp., “Standard Structural Arrangements,” NSRP,JUIY 1976,

24. “Guide for the Fatigue Strength Assessment of Tankers,” AmericanBureau of Shipping, June 1992.

25. Comstock, E., cd., “Principles of Naval Architecture,” SNAME, 1969.

26. Glasfeld, R,, Jordan, D., Kerr, M., Zoner, D., “Review of Ship StructuralDetails,” SSC-266, 1977,

27. Tanker Structure Cooperative Forum, “Workshop Report: Fatigue Life ofHigh Tensile Steel Structures,” 1991.

6-2

28. American Bureau of Shipping, “Rules for Building and Classing SteelVessels,” 1990, Paramus, New Jersey.

29. Jordan, CR., Krunpen, R. P., “Design Guide for Ship Structural Details,”SSC-331 , 1990.

6-3

.- --

(THIS PAGE INTENTIONALLY LEFT BLANK)

1

1+t

Appendix A

Analysis of Ship Structural Details

Used In Case Studies

L_.

-.. .—,

(THIS PAGE INTENTIONALLYLEFT BIANK)

-

--’”” .

A.1 CASE STUDY INTRODUCTION

The case studies presented below are used to illustrate the complex loading on shipstructure details, Linear Finite Element Analysis (FEA) was used to determine thefatigue design stress (m,) and resulting stress concentration factors (K=f). Theprincipal stress is used to characterize the stress and estimate stress concentrationfactors as described in this report. The stresses and details shown are applicationdependent and are used as a guide to develop the fatigue design strategy.

The following case studies are used to evaluate stress concentration factors

1) Double hull tanker frame cutout for a longitudinal and a deep bracket in atransverse frame.

2) Roll on-Roll off (Ro/Ro) ship side port

3) Double hull barge transverse floor cutout for a longitudinal.

4) Small Water Plane Twin Hull (SWATH) beam bracket in the haunch area,

A.2 CASE STUDY ANALYSIS

The first case study includes two details in a double hull tanker shown in Figures A-1and A-2. The midship section of the double hull tanker is shown in Figure A-3. Thisis representative of a mid size tanker (A-l). Hull loading for the double hull tankercase study is developed following the ABS Guide For Fatigue Assessment of Tankers(A-2). The structural loading developed using this guide is calibrated to a long termstress distribution parameter, Hydrodynamic loading for similar sized tankerspredicted by Bea, et al, (A-3) and Franklin (A-4) compares favorably with thepressure developed using ABS guidelines. The frame cutout and deep knee bracketare of interest because they experience fatigue failure (A-5). ABS guide recommendsfatigue analysis for both details (A-2). Typical frame cutout loading is shown inFigure A-4. Detail geometry and FEA models of the hull sections frame cutout and adeep knee bracket are shown in Figures (A-5) through (A-9). Stress concentrationfactors are shown in Tables (A-1 and A-2) for paneI stiffeners and deep brackets.

The Ro/Ro ship side port case study is of a detail common to Sealifl ships being builtin the United States (A-6). The Ro/Ro ship and side port are shown in Figures A-1 Oand A-1 1, The basic FEA model is shown in Figure A-12, Stress concentrationfactors are shown in Table A-3 for side cutouts,

The double hull barge case study is a cut out in the double bottom floor. Loadingand response data are presented by Fricke (A-7). The midship section and detail areshown in Figures A-13 and A-14, Stress in this cutout is shown in A-14.

r

The SWATH case study is a beam bracket in the haunch area of the strut. Loadingdata will be based on the data published by Sikora (A-8). Improved detail will bebased on the investigators knowledge of this type of detail in SWATH ships. TheSWATH ship, midship section and beam bracket are shown in Figures A-15, A-16,and A-17. The basic FEA model is shown in Figure A-18. Stress concentrationfactors are shown in Table A-4 for typical beam brackets,

It is interesting to note that the chocked beam bracket has the lowest & (1 .57).This must be compared to the ~ to fully understand evaluate its application. The t$for the weld between the bracket flange and beam flange is very important. The weldis loaded axially. 1$ for an axially loaded fillet weld is 5.5 and ~ for an axially loadedgroove weld is 2.63. Using the guidance provided in Section 4.1:

Groove weld & *K,ti=l .57*2,63=4.1

Fillet weld 1$ ● Kw = 1.57 ● 5,53 = 8.63

Clearly, the fillet weld has a high combined ~ and K,Cfat a. = .30y. For a plain beambracket

Fillet weld 1$ ● Kw = 3.0 ● 2.25 = 6.75

The plane bracket has a higher combined \ and K~Cfthan a groove welded flangebracket, but better than a fillet welded bracket for this application.

A-2

-

SIDEPLA

BOTTOM SHELL dPLATING

Figure A-1 Double hull tanker

A-3

. . . .

.- —-. ~----

—---———. --- ——--— ----- ——--

PROFILE

;________ +____-––__k-–-––--_k––_ --–-+ --------

/ TANK 5 \ TANK 4 i TANK 3 ! TANK 2 ! TANK 1 !+–---–––-+---–---–+--–--–-–-+-–--–---~----–-–-+

L________J_____-__J---_––--_’

PLAN

DWT 150,000 t

LBP 250 m

LOA 260 m

BREADTH 40 m

DEPTH 20 m

--, ,-, -, ,,, ,,CONSTRUCTION UUUbLL HULL

Figure A-2 Double hull tanker characteristics

A-4

Y

M,----// ./\\II

I I I I I I 1..1 I .+

I I I I&

/ .

t,, h”?~~l TTTTl~T TTITTT~l TTTITTTTLb’

r-l

a

,._

[

WEB FRAME

r SIDE SHELL

..L” .,, ,,--, .~

TORSIONJJFROM WEBFRAME

AND SHEAR FROMLATERAL LOAD(AXIAL LOAD INDECK & ❑OTTOM)

~ ~ENolNG INPLANE BENOING d

DOUBLE HULL TANKERCASE STUDY 1

LONGITUDINAL CUTOUT

Figure A-4 Loading on side shell frame cutout

A-6 L--

-7-

Figure A-5 Course mesh FEA model of double hull tanker

A-7

_——. .

d

“T(1.5t flange mm)

t2/3d

(30; rein) min 12

L

~L (1.5t flange max)

Figure A-6 Detail geometry of panel stiffener

A-8-

Figure A-7 FEA model of panel stiffener

T-

Table A-1 Stress Concentration Factors for Panel Stiffeners

A-10

-

LONGITUDINALBULKHEAD

a

R> Y4

TRANSVERSE

<

1! A

v v

BoTTOM SHELL

Figure A-8 Detail geometry for deep bracket

A-1 1

r

Table A-2 Stress Concentration Factors of Deep Bracket

ra

L... R= 94

KSCF=1 .52

A-13

Figure A-1 O Ro/Ro side port cutout

A-14

i--

A Av v

d

I

IB

A

I

_..__.._____––__–––_––;_––__– -

4 I<

<I

A Av v

ELLIPTICAL CORNER

Figure A-11 Ro/Ro side port cutout detail

A-15

Table A-3 Stress Concentration Factors for Side Port Cutout

l?h

KSCFCX2.0

R=300

I

RAOIUS CORNER

n-1--1‘s”””L I I

ELLIPTICAL CORNER

A-16

-

T-

Figure A-12 FEA model of side port cutout

A-17

.

.2

L-

ELUPSE Z 1 RAl10 OF

I MAJOR TO MINOR AXIS

50

w

—

– 25

>

--(

-+J ,

BULB FLAT

Figure A-14 Detail geometry for bulb plate cutout

A-19

Table A-4Stress Concentration Factors for Bulb Plate Cutout

LOCALPRESSURE:

NOMINAL

STRESS

DUE TO S. M:

STRUCTURAL

DETAIL AT:

INNER

BOTTOM

OUTERBOTTOM

ON WAVE CREST IN WAVE TROUGH

pxo P.1L5kNlm2

It$,,,l t,l, ttt,l+ll+llt Itt!lttt!ttt!t! lt!ttlftp= *5kN/m2P . 70kNlm2

Tn. 03 Ntmm2

rn, ,70 Nhm2

?n= -S8 Nlmm2

-n. :19 N(rn,n2

SCALE W U.: o Sw moo NIW”T7

O 73 145 KSI

A-20

___ -

Figure A-1 5 SWATH ship

A-21

:~

Figure A-16 SWATH ship midship section

A-22

L_-

SIDE SHELL

Figure A-1 7 Detailed geometry for SWATH ship beam bracket

Figure A-18 FEA model of beam bracket

A-24

Table A-4 Stress Concentration Factors for Beam Brackets

d

&F=2.32

d

~=2.27

A-25

A.3 REFERENCES

A-1

A-2

A-3

A-4

A-5

A-6

A-7

A-8

Chen, H.H., Jan, H.Y., ConIon, J. F,, and Liu, D., “New Approach fo[Design and Evaluation of Double Hull Tanker Structures, ” SNAMETransaction, 1993.

“Guide for the Fatigue Strength Assessment of Tankers,” AmericanBureau of Shipping, June 1992.

the

Schulte-Strathus, R., and R. G. Bea, 1993, “Fatigue Classification ofCritical Structural Details in Tankers: Development of Calibrated S-NCurves and System for the Selection of S-N Curves,” Report No.FACTS-l-1, University of California, Berkeley.

Franklin, P. and Hughes, O., “An Approach to Conducting TimelyStructural Fatigue Analysis of Large Tankers,” SNAME T& R-R41,September, 1993.

Exxon Corporation, “Large Oil Tanker Structural Survey Experience,”Position Paper, June 1, 1982.

Wood, W,, Edinberg, D., Stambaugh, K., and Oliver, C., “Prediction ofFatigue Response in TAK-X Side Port Structural Details,” Giannotti &Associates, 1982 (Proprietary).

Fricke, W. and Daetzold, H., “Application of the Cyclic Strain approachto the Fatigue Failure of Ship Structural Details,” Journal of ShipResearch, September 1987.

Sikora, J.P., Dinsenbacher, A., and Beach, J. E., “A Method forEstimating Lifetime Loads and Fatigue Lives for SWATH and Con-ventional Monohull Ships,” Naval Engineers Journal, ASNE, May 1983,pp. 63-85.

A-26

i_-

Appendix B

Development of Fatigue Notch Factors

——. ._. -

(THIS PAGE INTENTIONALLY LEFT BUNK)

-

B.O FATIGUE NOTCH FACTORS

B.1 DETAIL

Ship structural details vary in geometry and loading making it difficult to correlatethem to existing data developed for structural details in published literature. In orderto correlate the ship structural details geometries with test data, it is necessary todefine basic weld configurations that are, to the extent practical, independent of detailgeometry. The basic weld configurations associated with ship structural details canbe defined as:

1) Weld ripple of longitudinally loaded groove or fillet welds,2) Weld toes of transversely loaded groove welds,3) Weld toes of transversely loaded non-local carrying fillet welds,4) Weld toes of transversely loaded load carrying fillet welds and,5) Weld toes of fillet weld terminations.

Each of these five basic weld configurations and associated failure location iscorrelated to an equivalent detail from the fatigue data presented in Tables B-1 andB-2 from SSC-318 (B-1) and SSC-369 (B-2). The ~ values for each detail are alsoshown in Table B-1 and summarized in Table 3-1 for the basic weld configurations.

The definition of stress and K associated with the basic weld configurations is thenominal stress range as documented in Section 3.0 of this report. However, stressand 1$ associated weld termination common in ship details (shown in Table 4-1)requires re-evaluation to be generic in application. The weld termination associatedwith the straight attachment of detail 30 shown in Table B-2 is used to establish 1$ of3.6 at one plate thickness from the weld toe, FEA from the University of CaliforniaBerkeley (B-3) presents Kt at various distances from the weld toe for a pair of similarattachment details as shown in Figures B-1 and B-2. A stress concentration factorcan be inferred at one plate thickness from the weid toe. With this information, it ispossible to estimate ~ of 3.0. This is independent of attachment geometry, Becausethis new ~ is independent of attachment geometry, it can be used with stressconcentration factors associated with other detail geometries, This assumes that thedesigner has knowledge of the state of stress at one plate thickness from the weldtoe. This must be obtained from FEA of the detail or from the nominal stress and anassociated K,Cfas described earlier in this report.

B-1

—

;SC-.318V,ld,nm,

Dcmi13

IQIH

i ,All

I&f

8~

IO(G)

10Q

3(G)

1(0!9A

25A

31]

x

I?(G)

IOH4

69

IOM

16(G)

z7(n)

19

30A26

14II!1

7(P)

36

25n1216q

2 [(31$’I

‘ 202324

30

3817A17

18

WA27

3331A46

4032B

Table B-1

Fatigue Strength of Welded Details

>fcm Fmigu. s.cngth <M )., lE+l!6 Cycl.s ( k )

SC.318 AIIU. AIISY R.o R=(), Sy<5(, k,,

51 51.N 51 .

A83

46,5

38.3

39.242

36.1

31.231,341.s

10.9

38,1

30.3N

29.8

27.2

34

2s.32X3

25,7

25,2

23.62’!

24.3

17

13

17.1

m,

?I.n

20.4

206?0.619.6

19.919.2

18.1

16.117.217.1

16,716

15.6Is

11514.1

12

11,415.711.911.211.2

48.2

44.9

37,1

39.842.1

3s.2

31.531.2

3t,431.1

3S.8

2927,s

29,82,,2

3S,227.327.3

23.726,4

22.7

24.1

23.823.2

?317.4

25.9~~,,

11.H

21.5

2020

101

19.6

19.1

17.9

17.s18.318.3

16.7

16!6.2]4.6

12.3.14.112.8

11.615.611.911211.2

4S,6

42.1

36.2

39.141

32.832.7

31

38.428.829.3

29.1

27.1

28.427.1

33.126.826,8

25.8245

24.S

23.923.8

23.1

22.7

21.sxl

2019.7

!9.619,5

17.9

17.s

39338.1

362

35.4

3s31,6

.31

30.5

!9,729.6

29.?

28.328,1

27.?.

25.8

2s.73.5.7

2s5

245

245

24.S

24.4

23

23

12.9~?,,

:1.*

20

201,9.7LY.6

19,4

17.9175

16,716

I5,814.612.814.1

13.512.9

16,1

16

15.s14.614314.1

13312.9

15.8

....

S,axhrd U,vim,on .( Log 6.5( ksi uniu 1

U.o K=(l, Sy<311kv

U.U74

0.(!60.)040.04

0.094

0.076

0,136

0.114O.osd0.117

0.115

0.IG90.w9

0.055

0.G97

0,072

0.1020.CW2

0.092

0.079

0,093

0,215

Om0.0s3

0.1570.0!4

0.054

0.1150.018

{1,!17

0.075O.(M2(Km0.055

O.lcd0.045

0.0370.C99

0.(15I

0.058

0.05 I0,0460.107

0.0550,1010.0550.12

0.04

0.042

0.!34

0.079

0.017

0.127

0.0810.0s7

0.064

0.120.044

0.051

0.0450.072

0.101

0.095

0.095

0.0850.093

02150.08

0.11.

0.014

0.054

0,1090.080.111

.0.0620.1!420.05s

l).1U40.044

0.037

0,099

0.0s 1

0.0S8

0.0510.046

0.1480.0550.101

0.055

Kr

1.43.

1.43.

1.43.

1.43.1.54

1.43.

1,82

1.841.94

1,43.

2.CJ

2.052.072.15

2.11

2.16

I.s4

2.19

2.192.33

2,46

2.46

2.522.46

2.61

2,622.62

2.632.58?.(,9

2.73

3.11I~,y~

2.9s3.073.01

3.28

1.44

3,6

3.6.6

3.s14.264,74.164.46

4.67

3.71

F.,!$”. Cvmkinz(ixion SirU

\v,ld

Tcc

Weld

TceT,x

R{PPIC

T’x

\V.!d

TX

Ripple

Ripple

Rcet

Tcc

T%. or 1). T...T=

D. T.1,X

To?

TXT<.

Tm O, R,xnTcc

Trx

TrxTccT%

D, T

Tcc

u, r.D. T.

D. T.D. T,

Tc+ a, C.T. ., D. 1...T.xD.T.

TX and D. T.Tm and D. T.

‘P!ai. PM.o c. T.. Continuous Tmrhuuon. D. T.. Discomin.ous Tmnit!ati..

B-2 L

_-—

Table B-1

Fatigue Strength of Welded Details (con’t.)

ssc.318 Ma Fauwe Sum@( AS ) al lGW C@m ( M ) sumda@Devmaucnof l..% AS u Fat>#neCrockWdd,mmt ~=o ( ti, u.is)

mm!,Imitiadm SiIca

SSC.118 A1l R, AIISY R=O R= O, S,< Ylks, R.O. Sv<50ksi21(s) 31 31 305 105 0.U31 0.031 I ,97 Tce18(S) 20 21 21 0.041 0.042 2.8733(s) 2?’ 203 20.7

T.x d D. T.20.7 0.% 0.!3s 2.91 Tm

17(s] 11 21 19.6 19.6 O,cdl 0,041 3.07 Toe17MS) 21 21 19.6 19.6 0.041 O,U1 3.07 Toe2CW) 19.6 ?.12 16.9 173 0.159 0.163 356 Tot19(s) 20,3 182 15.4 15.4 0.124 0.124 3.91 T-3$(s) 13 [3.3 133 133 0.113 0.113 ,46 Tm

B-3

. . .. .—

iTEGORY

A

B

DETAILNUMSER

1

2

8

l(F)

Table B-2

Welded Detail Classification

DESCRIPTION ,LOADING

lain plate,achined edges,xial

Lolled I-Beam,3ending

Double shearlap joint,Axial

bolted

PICTOGRAPH

‘=ak

--=

Plain plate flame-cut edges, Axial -Qsk

KWIOSYIIWLS kpresentedonPageB-16 B-4

Table B-2

Welded Detail Classification (con’t.)

:ATEGORYDETAIL DESCRIPTION,NUMBER LOAOING

PICTOGRAPH

Longitudinally3 welded plate, as-

welded, Axial‘*.

(As-welded)

Longitudinally3(G) welded plate, weld

ground, Axial ‘*.

(Groundfacesoftheweld)

B

Transverse butt1O(G) joint, weld ground,

Axial

(~,

f.l%ldfac=~und)

Transverse butt10A joint, as welded,

M%)In-plane bending(k-welded)

Key to symbols iSPresented on Page B.16 B-5

Table B-2

Welded Detail Classification (con’t.)

,V17PARV DETAIL DESCRIPTION,--.--..” PICTOGRAPH

,------- NUMBER JJUIUIJ.NW 1

25ALateral attachmentto plate, Axial

‘%’

Flange spliCeB 13 (unequal width), (~)

as-welded, BendingSqa-u.l

(As-welded)

28plain plate with

drilled hole, Axial ‘--

(Drilled hole)

Flange splice

c 12(G)(unequalthickness), weld (*)

ground, Bending W==.I

(WeldfacesgrOd)

Key to symtds is presented 0. Page B-16B-6

—. —..

I

Table B-2

Welded Detail Classification (con’t.)

Key m wtnbls is presented. . Page 8-}6 B-7

,—--- ._

Table B-2

Welded Detail Classification (con’t.)

\TEGORYDETAIL DESCRIPTION, PICTOGRAPH.,.wmmm T.i3an TNG.“”=. - ---- . ..-

25Lateral attachments ‘to plate, Axial

%

c_

c

I-beam with welded

(a

I

7(B) stiffeners, Bendingstress in web

Qc

3OALateral attachmentsto plate, Bending [e,

)

D

Doubler pla’Ce26 welded to plate,

Axial ‘--

Kqtosymbk ispresented on Page B-16B-8

._ .—

Table B-2

Welded Detail Classification (con’t.)

ATEGORYDETAIL DESCRIPTION,NUMBER LOADING

PICTOGRAPH

14Cruciform joint,

‘%-

CAxial

Transverse butt

11 welded I-beam, as- (a)

welded, Bending

(.%-welded)

D

Cruciform joint,

‘%)

G

1/4” weld, In-plane !21 bending stress at i---c

weld toe, C I

I-beam with weldedstiffeners,

[a1

7(P) Principal stress inweb

)c

Key to symbols ispresented on Page B-16 B-9

_,—......-—

Table B-2

Welded Detail Classification (con’t.)

:ATEGORYDETAIL DESCRIPTION ,NUMBER LOADING

PICTOGRAPH

Welded beam withintermittent welds

[

,36 and cope hole in

the web, Bending LP

-)

Lateral attachment ‘25B to plate with

stiffener, Axial

D

Flange Splice

12(unequal (=)thickness), as- %-==.1welded, Bending

(As-welded)

Partial penetration16 butt weld, as-

welded, Axial ‘-.

(Partialpenetion - as-welded)

-

Keytosymbds ispresented on Page 516 B-1 O

:ATEGORY

D

E

DETAILNUMBER

22

21(3/8”)

20

23

Table B-2

Welded Detail Classification (con’t.)

DESCRIPTION ,LOAD ING

\ttachment of studzo flange, Bending

Cruciform joint,3/8” weld, Bendingstress on throatweld

Cruciform joint,Axial, Stress onplate at weld toe C

Attachment ofchannel to flange,Bending

Key JO s’w)L351S is presented on Page B-18 B-1 1

<

PICTOGRAPH

)—

--- .—.

Table B-2

Welded Detail Classification (con’t.)

Key to syrnbds is presented on Page B-16 B-12

+—’---

Table B-2

Welded Detail Classification (con’t.)

:ATEGORYDETAIL DESCRIPTION,NUMBER LOADING

PICTOGRAPH

Channel welded to17A plate, longitudinal

weld only, Axial-%

Attachments of31A plate to edge of

flange, Bending (w,

F

Angles welded onplate, longitudinal

17 welds only, AxialStress in angle end ‘b?of weld, C c

Flat bars welded toplate,

18 longitudinal weldonly, Axial Stressin plate, C -&-

Key to symbols is presented on Page B-1.5 B-1 3

,

-——. .——

Table B-2

Welded Detail Classification (con’t.)

ZATEGORYDETAIL DESCRIPTION,NUMBER LOADING PICTOGRAPH

F 3.2A:~gd::;,e (-)

flange at end ofattachment, C

Slot or plug welded27 double lap joint,

Axial ‘%=

(S10[or Plug Welds)

Flat bars welded to

G 33plate, lateral andlongitudinal welds,Axial -%.

Triangular gusset46 attachments to

‘B-

C

plate, Axial

Key. to symbols ,sQresentedo” Page B.16 B-14

L-

Table B-2

Welded Detail Classification (con’t.)

?ATEGORYDETAIL DESCRIPTION,NUMBER LOADING

PICTOGRAPH

7’

-

(

e

c

Interconnectingbeams,

40Bending in

perpendiculardirections

c

)

G L

Butt welded flange32B (unequal width),

Bending (-)

‘%)

OCruciform joint, /

21(s)In-plane bending, LoShear stress on the 1weld, C.

s

Flat bars welded toplate, longitudinal

18(S) weld only, Axial,Shear stress on ‘&-weld, C~

Key 10 symbols is presented on Page B-16 B-15

Table B-2

Welded Detail Classification (con’t.)

WIEGORYDETAIL DESCRIPTION,.“-. .. 7nnnrnm

PICTOGRAPHm“,.sc.n ..,”C.” . ..-

Flat bars welded toplate, lateral and

33(s) longitudinal welds,Axial, Shear stress

‘*.on weld, C~

Angle welded toplate, longitudinal

17(s) weld only, Axial,Shear stress onweld, C. ‘%.

s

Channel welded toplate, longitudinal

17A(S) weld only, Axial,Shear stress onweld, C~ ‘%

Cruciform joint,20(s) Axial, Shear stress

-%-

0

on weld, C$

Key 10 symbols is presented on Page B-16 6-16

-

Table B-2

Welded Detail Classification (con’t.)

Key to Symbols

(F) -(G) -(B) -(P) -(s) -A,B,C, .c+ -c~+ -L-P-R-t-

Flame cut edgesweld groundBending stressesPrincipal stressesShear stressesAdditional description within the same detail numberCrack initiation site due to tensile stressesCrack initiation site due to shear stressesLength of intermittent weldPitch between to intermittent weldsRadiusThickness of plate

B-17

— FINE MESH

— STANDARD MESH

1.7 T

1.6

1.5

1.4

~ 1.3

1.2

1.1

1

0.90 20 40 60 80 100 120 140

DISTANCE FROM HOT–SPOT (MM)

Figure B-1 FEA of attachment detail I

B-18I

~.......-

EEEG1.8

1.7

1.6

1.5

1.4

+ 1.3

1.2

1.1

1

0.9

0 20 40 60 80 100 120 140

OISTANCE FROM HOT–SPOT (MM)

Figure B-2 FEA of attachment detail II

B-19

——. .—1-

6-2 REFERENCES

B-1

B-2

B-3

Munse, W. H., T, W. Wilbur, ML, Telalian, K. Nicol, and K. Wilson, 1983,“FatigueC haracterizationo fFabricated Ship Details for Design,’’ ShipStructure Committee Report SSC-318, Washington, DC,

Stambaugh, K., Lesson, D., Lawrence, R., and Banas, “Reductionof S-NCurves for Ship Structures,” SSC-369, for Ship Structures Committee,1992.

Schulte-Strathus, Rand R.G. Bea, 1993, “FatigueC lassificationo fCriticalStructural Details in Tankers: Development of Calibrated S-N Curves andSystem forthe Selection of S-N Curves,’’ Report No. FACTS-l-1,University of California, Berkeley,

B-20

-

Appendix C

Analytical Fatigue Life Prediction

“r

i

““

I

~_._... —.—.

C.O INFLUENCE OF WELD PARAMETERS ON FATIGUE LIFE

The use of & for the weld termination and other weld configuration permits relatively

easy comparisons between details and associated stress concentration factors.

Thus, the K, approach for basic weld configurations includes the effects of suchimportant factors as weld geometry, residual stress and mean stress because it isbased on test data of actual welds. This assumes good welding practice which issomewhat subjective and produces much of the unaccounted for scatter in the testdata. The following analytical expressions are useful to determine the impact ofcontrolling these parameters.

C.1 ANALYTICALLY ESTIMATING AS WELD

From Basquin’s Law, Yung, and Lawrence (C-1) derived an expression to calculatethe mean fatigue strength of weidments at long lives:

2 (a; u) (2N,)bAswe,d=

Keff

[

,+I+R(2N,)b

1

(c-1)fmm

I-R

where:

U’f = Fatigue strength coefficientUr = Local (notch root) residual stressb = Fatigue strength exponentR = Load ratio

N, . Cycles to failure = - J[1

log 2+ ~ (MPa , mm )

effK=fmm Effective fatigue notch factor = ( 1 - X ) K~.x + X K~a.

c-1

saT= seA+ s:

In the following section, we have assumed that the total fatigue life (NT) is equal tothe initiation fatigue life (Nl), It is expected therefore that this estimate of fatiguestrength will be accurate only at long lives ( = 107 cycles) for which the propagationportion of the fatigue life ceases to be important.

C.1.l Ri@e

The idea is that the fatigue notch provided by the ripple can be treated as an infinitearray of semicircular notches. The weld metal properties determine the fatigueresistance and the residual stresses in the as-welded state.Thus:

K, = 1.6

K fmex = [1K,-1l+— = 1.3

2.

C+f,b = f( SUwm)

CTr = SYWM

Lc-2

2SUWM- 2SYWM+ 689.5ASwe,d =

1.30

C.1.2 Groove Welded Butt Joint

~. .-

2N~

[11

(C-2)1+ ~ 2Nf

I-R

The idea is that the residual stresses at the toe of the groove weld are controlled bythe yield strength of the base metal and that crack initiation occurs in the HAZ;therefore the HAZ properties control the fatigue resistance. Thus, for the as-weldedstate:

~’zb = f(SuHAZ)

u, = S,,M

\:ax = 1 + 0.0015(.27) (tan 8)02’ SUHAZ m (MPa, mm)

K:ax = I + 0.0015(0.165) (tan f3)0”7 SuHAZfi (MPa, mm)

2SUHAZ 2SY,U + 689.5ASwe,d=

Keffmax

2NI b

1+

[11~ 2N:

(c-3)

C.I.3 Non-Load Carrvim Fillet Weld

The idea is that the residual stresses at the toe of the fillet weld are controlled by thevield strenath of the base metal and that crack initiation occurs in the HAZ and~herefore t~e HAZ properties control the fatigue resistance, that is all is as in B.2.2above except for differences in the models for \ which include the effect of the LOP(2c) oriented parallel to the applied stress:

~’f , b = f(SuHAZ)

u, = SVBM

c-3

~ ~.

Case 1- A model for K,~.Xwhich considers the effect of LOP

[ [v651suHAzf~ax=1 + 0.0015 (.35) (tanO)025 1 + 1.1K (MP,,mm)

K&,x = 1 + 0.0015 (.21) (tan8)01e7 Su~Az J (Mf,, mm)

Case 2- A model for ~.,, which considers the effect Ieq Ierwth

[K& = 1 + 0.0015(.04) 2 - ; SUHAZ~ (MP~, mm)

2Su~~z - 2SY,~ + 689.5Aswe,d =

II

2N:(c-4)

k;:x 1+

[1

~ 2N;I-R

C.I.4 Load Carrvinq Fillet Weld

The idea is that the residual stresses at the toe of the fillet weld are controlled by theyield strength of the base metal and that crack initiation occurs in the HAZ and;therefore, the HAZ properties control the fatigue resistance, that is, all is as in C. 1.3above except for differences in the models for Kf which include the effect of the LOP(c) now oriented perpendicular to the applied stress:

~.X = 1 + 0.0015 (.35) (tan8)025K[1 ‘1” H“51SUHAZ f ‘Mpamm)

c-4L

-

K&,X = 1 + 0.0015 (.21) (tanL3)0”7

Load Carrvinq Fillet Weld: Root Failure

u’,, b = f(SuM)

~.o

s“HA, F

K-frnax ~

K-fmax B

Aswe,d

[105

1 + 0.0015(1 .15)(tanO)-02 ~ SUW~ t

, . 0,0098 ~ O’z s“Wt’1 fl

l+;/

1w’–[1

2

2;2 Ct’

2S”WM+ 689.5

i

2N:

Keflfmax 1+

[1@ 2N)I-R

(Mf,, mm)

(MP~, mm)

(MPa, mm)

(c-5)

At Ions lives, the likelihood of LOP failure in a weldment is increased. Increasingplate ~ickness (t) increases the tendency for LOP failure

C.I.5 A Fillet-Weld Termination

For fillet weld terminations, the residual stresses at the toe at the end of the filletweld are controlled by the yield strength of the weld metal and that crack initiationoccurs in the HAZ and therefore the HAZ properties control the fatigue resistance.The models for K, include the three dimensional effects of flow of the stress in themain plate into the weld and the attachment. This quantity is captured as a stressconcentration factor of SCF which is the ratio of the stress at the location of thehypothetical strain gage a distance (t) away from the toe of the weld to the stresses

c-5

~

at the station of the weld toe in the absence of the weld. These results must bedetermined from the results of the original FEA

a’,, b = f(SuHAZ)

u, = SYm

SCF = 1.5

K~eX = SCF*[I + 0.0015(,35)(tan8)025~ 165

I + 1.111

s.... 11 (~pa mm)i

K’ - SCF*[I + 0.0015(.21 )(tanr3)0167 SUH*Z []f,na,

M&,, =

‘“x [+[%4

3.0 SU,ti - 1.55SUw~ + 689.5

(MPa, mm)

(C-6)

C.2 EFFECTS OF BENDING

Bending of attachments on plate is important because of minimal section depth.Bending of the plate causes stress gradient effects. Only one of the weld details inTable B-1 was subjected to pure bending. All others, while subjected to a nominalbending load, are of such a depth that the stress state at the fatigue initiation site isfor all purposes an axial load, thus the loading is considered pseudo-axial. However,from this one example comparing SSC-318 (C-2) detail 30 with 30A shows that therecan be a large difference to the fatigue response of weldment to pure axial and purebending loading (ASd.,l~n = 99.6 MPa, axial, ASd.~igfl= 144.3 MPa, bending). Thiseffect is captured by the analytical expressions for AS’weldand as well as by theexperimental database; however, there are very few pure bending entries in Table B-1 probably because the data has been restricted to R = O loading conditions.

C-6 L

-~

The analytical expressions of AS weld can deal with various combinations of axial andbending loads directly or provide an expression for predicting the expected meanfatigue life at a given long life using the ex~erimental results for pure axial and purebending loading:

AS:.,. * A :.,6AS$; =

AS:e,, (1 - X) + AS:.,, (X)

(c-7)

where:

AS:.,, = Experimental fatigue strength under pure axial loading

AS:.,~ = Experimental fatigue strength under pure bending loading

The weld termination represents a large challenge because it cannot be dealt withadequately using a 2-D stress analysis. If the situation were axial loading, the ratio ofthe stress at the location of the hypothetical strain gage a distance (t) away from thetoe of the weld to the nominal stresses at the station of the weld toe would be 1.However even in 2-D states of stress, local bending can cause stress gradientsindependent of the stress-concentrating effects of the weld toe. In situations such asthe weld termination the relationship between the nominal stress at the location of thestrain gage and that at the station of the weld toe is dependent upon many factors.To solve this problem the designer determines the stress at the location of the weldtoe from the results of a finite element method and expressed it as an SCF.Incorporation of this SCF into the expression for 1$ for the fillet weld and assumingthe high level of tensile residual stresses possible because of the shrinkage of theweld metal, leading to the creation of an analytical model which predicted thebehavior of the weld termination. It is believed that this process can also be used forother weld shapes having a geometry which cannot be analyzed as a simple “2-D”FEM problem.

Under pure axial loading and for normal weld toes (O < 450), fatigue is predictedalways to occur at the root. Under pure bending loads, fatigue failure will alwaysoccur at the toe before the weld root. Note that most axially loaded welds haveinduced bending stresses at the weld toe due to the straightening of weld distotilonsunder axial load. This effect induces secondary bending stresses which can easilycause the weld toe to become the failure site even under nominally axial loadingconditions.

In Figure C-1 the ratio of ASW,ld(root)/ASW,, $toe) is plotted against the ratio of bendingstress amplitude to total stress amplitude (x). As seen in Figure C-1, when the value

c-7

.

T%.%..;●..

“-8%*%i

‘..“.

“.●

07

:2

I i:::

.................................... ....

....... ........................ ..........

m,

“.,

\

. ...\ti ........................

●.

‘f.,“.“~‘.%t\%“.:

.................... ............

g’-

es

..._.....&.. ..............

?2&

C-8L-

of (x) exceeds about 0.3 (that is, when the ratio of bending to M stresses is above0.35, the failure location should shit? to the weld toe. Unfortunately the most effectiveway of improving the fatigue life of the load-carrying fillet weld failing from its root isto increase the weld penetration, that is, to reduce the value of (c). This change hasa large beneficial effect upon ASW,[~(root), but also improves the performance at theweld toe so that shifting the failure location from the root to the toe requires largereductions in the value of (c). Note again that the above discussion assumes a zerovalue for the welding residual stresses at the weld root.

Calculations were made using the Initiation - Propagation Model to approximate thetotal fatigue life. The initiation life calculations was slightly altered to take intoaccount the set-up cycle. The propagation life calculation was made usingexpressions for M~ (C-4).

C.3 CALCULATIONS MADE USING THE ANALYTICAL EXPRESSIONS

Calculations were performed forhot-rolled steel undera load ratio (R)=O. From thework of McMahon and Lawrence (C-7), the relationships between the ultimatestrength of the base metal and the ultimate strength of the heat affected zone and theyield strength of the base metal (Figures C-l and C-2) forhot-rolled steel were foundto be:

s YBM = (5/9) *suBMs “HZ = 1.5* S”BM

Areasonable (assumed) relationship between theultimate strength of the weld metaland the yield strength of the weld metal was assumed to be:

s y!ml = (7/9) ’suw

In addition, the following values were assumed:

:=

t=

suBM =

s “m =s yBM =

s yHAZ ‘

s yb’vM =

0.045°19mm414 MPa483 MPa(5/9) * SU,. = 230 MPa1.5’ SU,~= 621 MPa(7/9) ● SUW = 376 MPa

Theresults are plotted in Figures C-3through C-8together with themean S/N data ofthe Iocalfatigue details from SSC-318 (C-2). In Figure C-3, the analytical expressionfor the ripple primitive underestimates life but is very reasonable and a somewhat

c-9

xCL(3

N<r

600

500

40C

30C

20C

I 0(

(

mm

/

HHAZ=1.5 HBM o-

V

1

4/loo

$

p

0

0

/m

-/

0 ChongA Vorious.9ructurol SteelsoASTM A36O ASTM A514

❑ o Loser DressedU ~ T IG Dressed

/

a ASTM A36, HighHeot Input

II Ronge of HA2 Hordness

100 200 300 400

Base Metal Hardness, DPH

Figure C-2 Hardness of the heat affected zone as a function of base metalhardness.

2000

180C

160[

140(

.&!z=- 120(

“:

wLG I00(-clz.—>

80(

f5(3(

40(

20

Su , ksi

~

~o j❑ HoI RolledO Quenched ond Tempered

8 o/A Cold DrownO Normalized

/

o

E ● Filled Symbols: HAZ- CUOOChong o

0’0/

●

TemperedM

o~-o — Quenched and

F/

/

00I I 1 I I

100 200 300 400 500

Hardness, DPH

1 I I I I

o 400 800 1200 1600 2000

)0

1

SU , MPa

Figure C-3 Yield strengthas a functionof hardness

c-1 1

,,!

HT

I

I

I

I

0

........... .... .. .. ...... .

,.. r

-$?-0cm

c-1 2-

_. ._— -.–

. . . . .

I

1/t--+!1

-1

.

‘o

00

g

-0$=C3mm

100

At-“’-------... ... . .. ... .. . .. .----------... . ... . .. ... .. ... ... . .... .

()

10 . ........................................................................................................................... .

~ Detail#25

.. ..~.--~tail #25A

_ I, Primitive F - Non-ti CW~gFillet Weld, Case 1

1---------- . . . . . . .. . . . . . . . . . . . .-

“-- .”-----.......................AL

<)

Fatigue Life, N (cycles)

I

Figure C-6 Comparison of the S-N data for ship structure details and the analyticalexpression for the weld Primitive F - Non-Load Carrying Fillet Weld

100

10

,

~ Detail#25

. ..A--- Detail #25A

— I,,PrimitiveF - Non-LoadCarryingFdlet Weld, Case 2

.- . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .. .

-- ”.-.-....................

ltf106

,.7

FatigueLife,N (cycles)

Figure C-7 Comparison of the S-N data for ship structure details and the analyticalexpression for the weld Primitive F - Non-Load Carrying Fillet Weld

100

10

!

~ Detail #14...@.. Detail #20+ - Detail #21—... - Detail #21(3/8)— & - Detail #26— I, Primitive F - Load Carrying

Fillet Weld1

I

lC$ ~06 ~07

FatigueLife,N (cycles)

Figure C-8 Comparison of the S-N data for ship structure details and the analyticalexpression for the weld Primitive F’ - Load Carrying Fillet Weld

r

,—

conservative approximation. In Figure C-4, the analytical expression for the groovewelded butt joint primitive is in very good agreement with the S-N data. It isespecially good at very long lives when initiation dominates fatigue life and theabsence of the propagation life does not make much of a difference, The analyticalexpressions for the non-load carrying fillet primitive are reasonable as shown inFigures C-5 through C-8. The analytical expressions for the termination primitive isalso reasonable at long lives but is too conservative at shorter lives (N<l E+06).

The results are plotted in Figures C-9 through C-12 together with the mean S-N dataof the local fatigue details from SSC-318. No calculations were made for the ripple(R) and groove weld (G) primitives because values of Mk were not available. The I-Pcalculations for the non-load-carrying fillet primitive are reasonable as shown inFigures C-9 and C-1 0, Comparing the I-P calculations with the I calculations, onecan see the significant effect of propagation at shorter lives and its almost negligibleeffect at longer lives. In Figure C-11, the I-P calculation for the load-carrying filletprimitive is in good agreement with the S-N data. An increase in fatigue life is seenat high stresses due to the addition of the propagation life; but once again, not muchof a difference is seen at long lives. The termination calculations agree with the S-Ndata over all the lives due to the addition of the propagation life but the estimatedinitiation portion of life seems to be a bit too long (un-conservative).

c-1 7

500

r

I300

‘h.

40

20

.

r-

~ Detail#17..=-- Detsil#17A--e - Detail #18+- - Detail #30..-+.. - Detail #32A~ Detail #33— I, PrimitiveT - Weld Termination

:.>::_~--:~-...-''''.''' .'.'..''''.''-''''-'''..''.'.''.''.'...'..''''-'-""-"""""

~

,

[($ 106

Fatigue Life, N (cycles)

Figure C-9 Comparison of the S-N data for ship structure details andexpression for the weld Primitive T - Weld Termination

the analytical

_,. r--------

500

300

‘g

40

20

!

~ Detail#25

... .EI..- Detail#25A

_ 1P,PrimitiveF -Non-LoadCarryingFilletWeld,Case1

“--------- ---------- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,- ..6-..-.-...

“-- . . . . . . . . . . . . . . . . . .....

““”-------....................,

106

1 i

107

Total Fatigue Life, NT (cycles)

Figure C-1 O Comparison of the l-Pmodel predictions and S-N data of ship structuredetails fortheweld Primitive F- Non-Load Carry ing Fillet Weld LoadCase 1.

‘“[

,

oc+

_ .— _

: ,

tu

&-

I........................- ..... “t). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

L-

~.

C.4 SUMMARY

The predictions of fatigue strength using the analytical expressions given for each ofthe primitives agree with the experimental results at long lives (N = 107). Thus, theexpressions are able to predict primitive behavior through are able to predict primitivebehavior through a knowledge of the weldment material PrOp@ieS (S.), residualstress (or), loading conditions (x), plate thickness (t), and weld 9eometry (8, 1,c).Finally, analytical predictions of the fatigue behavior of the primitives made using theI-P model were good at both long and short lives. Thus, it would appear that the I-Pmodel agrees well with the experimental results but has the powerful advantage ofrevealing to the engineer the true importance of interconnectedness of the manyfatigue variables influencing the fatigue behavior of a given primitive (weldment).

By assigning average values to the fatigue parameters reflected in the data baseinformation for a primitive, the designer can gauge the anticipated effect onexperimental S-N diagram information of substantial changes in: weldment size (t), Rratio, loading conditions (x), base metal and weld metal strength (Su, SY), weldgeometry (8, 1,c), residual stress conditions through the use of the expression belowand the provided analytical expression for the app~opriate primitive.

1.43 AsppM-weld=[

A.SW,,, calculated for current conditions

‘CF * ‘hwld ‘Sweid ca/cu/ated for standard conditions

C-23

C-5 REFERENCES

C-1 Yung, J.Y. and F.V. Lawrence, 1985, “Analytical and Graphical Aids forthe Fatigue Design of Weldments,” Fatigue Fract. Engineering Mater.Struct., Vol. 8, No. 3, pp. 223-241.

C-2 Munse, W. H., T. W. Wilbur, ML. Telalian, K. Nicol, and K. Wilson, 1983,“Fatigue Characterization of Fabricated Ship Details for Design,” ShipStructure Committee Report SSC-318, Washington, DC.

C-3 McMahon, J.C. and F.V. Lawrence, 1984, “Predicting Fatigue PropertiesThrough Hardness Measurements,” Report of the Materials Engineering -Mechanical Behavior, University of Illinois at Urbana-Champaign.

C-24

,.

L---

APPENDIX D

Glossary

(THISPAGE INTENTIONALLYLEFT BUNK)

Cathodic protection

Constant amplitudefatigue limit

Continuoustermination

Cruciform ortransverse load-carrying joint

Cut-off limit

Design life

Detail category

Discontinuity

Discontinuoustermination

A means of reducing corrosive attack on a metal bymaking it the cathode of an electrolytic cell. This canbedone by applying anexternal direct current fromapower source (impressed) or by coupling itwith a moreelectro-positive metal (sacrificial).

The fatigue strength at 5-106 cycles. When all nominalstress ranges are less than the constant amplitudefatigue limit for the particular detail, no fatigueassessment is required.

Termination from continuous weld

Specimen made from two lengths of plate welded, viafillet or full penetration welds, to either side of aperpendicular cross piece of the same sectionthickness.

The fatigue strength at 108 cycles. This limit iscalculated by assuming a slope corresponding tom = 5 below the constant amplitude fatigue limit. Allstress cycles in the design spectrum below the cut-offlimit may be ignored unless the detail is exposed to acorrosive environment.

The period during which the structure is required toperform without repair.

The designation given to a particular structural detail toindicate which of the fatigue strength curves should beused in the fatigue assessment, The category takesinto consideration the local stress concentration at thedetail, the stress direction, and residual stresses.

An absence of material causing a stress concen-tration.Typical discontinuities are cracks, scratches, corrosionpits, lack of penetration, slag inclusions, cold laps,porosity, and undercut.

Termination from intermittent weld

D-1

Fatigue The damage of a structural part by gradual crackpropagation caused by repeated stresses.

Fatigue limit

Fatigue loading

Fatigue design stress The stress in a structural member at the location of theweld and at one plate thickness from theweld toe forweld termination typical for calculated using FEA. Thisstress iscorrelated to nominal stress range todetermine fatigue life.

See “cut-oft” limit.

Fatigue loading describes the relevant variable loadsacting on a structure throughout the design life. Thefatigue loading in ships is composed of different loadcases.

Fatigue notch factor Ratio of stress of a notched detail to stress for a plandetail at a constant fatigue life.

Fatigue strength The stress range corresponding to a number of cyclesat which failure occurs.

Hot spot stress

Geometric stress The stress at any point around the detail inter-sectionnecessatyto maintain the compatibility ofdisplacements. This stress excludes local Streisanddepends on the nominal stress and overall geometry ofthe intersecting members.

The stress which controls fatigue endurance in tubularnodal joints. It can be defined experi-mentally or indesign by the product of the nominal stress and thedesign hotspot stress concentration factor. This formis used primarily for offshore structural details.

Load case

Load stress

A part of the fatigue loading defined by its relativefrequency of occurrence as well as its magnitude andgeometrical arrangement.

The stress due to the discontinuity at the weld andwhich is superimposed on the geometric stress.

D-2

L-

.T-

Nominal stress The detail stress remote from the intersection. This

includes geometric stress at the weld toe in theabsence of weld.

Nominal stress range The algebraic difference between two extremes

(reversals) of nominal stress. Usually, this difference is

identified by stress cycle counting. Stress extremes

may be determined by standard elastic analysis and

applying forces and moments to the cross-sectional

areas. Exceptions to this definition are details near cut-

outs, man-holes, or other stress concentrations not

shown in Table 3-1.

Ripple

Weld profiling

Weld toe

Uneven weld surface

Process of mechanically altering weld surfacegeometry,

The intersection of the weld profile and parent plate.

D-3

L--

-— . ..r

Project Technical Committee Members

ThefollowingpersonsweremembersofthecommitteethatrepresentedtheShipStructureCommhtee totheContractorasresidentsubjectmatterexperts.As suchtheyperformedtechnicalreviewoftheiritialproposalstoselectthecontractor,advisedthecontractorincognizantmatterspertainingtothecontractofwhichtheagencieswereaware,andperformedtechnicalreviewoftheworkinprogressandeditedthefinalreport.

ChouL]rr- Chairman

GaryNorth

SteveArntson

PaulCojeen

Jaideep.Sirkar

MikeSieve

DaveIGhl

RobertClark

W. WilliamSiekierka

Mr. Alex StavovyDr.RobertSielski

CDR SteveSharpe

MaritimeAdministration

MaritimeAdministration

AmericanBureauofShipping

U.S. Coast Guard

U.S.CoastGuard

NavalSeaSystemsCommand

NavalSurfaceWarfareCommand

Clark-CimInc.

NavalseaSystemsCommand,ContractingOfficer’sTechnicalRepresentative

NationalAcademy ofScience,MarineBoardLiakon

U.S.CoastGuard,ExecutiveDirectorShipStructureCommittee

COMMllTEE ON MARINE STRUCTURES

Commission on Engineering and Tachnicai Systems

Nationai Academy of Sciences - National Research Councii

The COMMilTEE ON MARiNE STRUCTURES has technicai cognizance over the

interagency Ship Structure Committee’s research program.

Peter M. Paiermo Chairman, Alexandria, VA

Subrata K. Chakrabarti, Chicago Bridge and iron, Piainfield, iL

John Landes, University of Tennessee, Knoxviiie, TN

Bruce G. Coiiipp, Marine Engineering Consultant, Houston, TX

Robert G. Kiine, Marine Engineering Consultant, Winona, MN

Robert G. Loewy, NAE, Rensselaer Poiytachnic institute, Troy, NY

Robert Sielski, Nationai Research Councii, Washington, DC

Stephen E. Sharps, Ship Structure Committee, Washington, DC

LOADS WORK GROUP

Subrata K. Chakrabarti Chairman, Chicago Bridge and iron Company, Piainfieid, iL

Howard M. Bunch, University of Michigan, Ann Arbor, Mi

Peter A. Gaie, John J. McMuiien Associates, Arlington, VA

Hsien Yun Jan, Martech incorporated, Neshanic Station, NJ

John Niadzwecki, Texas A&M University, Coiiege Station, TX

Soiomon C. S. Yim, Oregon State University, Corvaiiis, OR

Maria Ceiia Ximenes, Chevron Shipping Co., San Francisco, CA

MATERiALS WORK GROUP

John Landes, Chairman, University of Tennessee, Knoxviiie, TN

Wiiiiam H Hartt, Fiorida Atiantic University, Boca Raton, FL

Horoid S. Raemsnyder, Bethiehem Staei Corp., Bethiehem, PA

Barbara A. Shaw, Pennsylvania State University, University Park, PA

James M. Sawhiii, Jr., Newport News Shipbuilding, Newport News, VA

Bruce R. Somers, Lehigh University, Bethlehem, PA

Jerry G. Wiliiams, Conoco, inc., Ponca City, OK !1I

‘L_

SHIP STRUCTURE COMMllTEE PUBLICATIONS

SSC-356

SSC-357

SSC-358

SSC-359

SSC-360

SSC-361

SSC-362

SSC-363

SSC-364

SSC-365

SSC-366

SSC-367

SSC-368

SSC-369

SSC-370

SSC-371

SSC-372

SSC-373

SSC-374

None

Fatigue Performance Under Multiaxial Load by Karl A. Stambaugh,Paul R. Van Mater, Jr., and William H. Munae 1980

Carbon Equivalence and Weldability of Microalloved Steels by C. D.Lundin, T. P. S. Gill, C. Y. P. Qiao, Y. Wang, and K. K. Kang 1990

Structural Behavior After Fatigue by Brian N. Leis 1987

Hydrodynamic Hull DamDing (Phase 1)by V. Ankudinov 1987

Use of Fiber Reinforced Plastic in Marine Structures by Eric Greene1990

Hull Strapping of Ships by Nedret S. Basar and Roderick B. Hulls 1890

Shipboard Wave Heiaht Sensor by R. Atwater 1990

Uncertainties in Stress Analvsis on Marine Structures by E. Nikolaidisand P. Kaplan 1991

Inelastic Deformation of Plate Panels by Eric Jennings, Kim Grubbs,Charles Zanis, and Louis Raymond 1991

Marine Structural Integrity Proarams (MSIP) by Robert G. Bea 1992

Threshold Corrosion Fatiaue of Welded Shipbuilding Steels by G. H.Reynolds and J. A. Todd 1992

Fatigue Technolocrv Assessment and Strategies for Fatigue Avoidancein Marine Structures by C. C. Capanoglu 1993

Probability Based Ship Desian Procedures: A Demonstrationby A. Mansour, M. Lin, L. Hovem, A. Thayamballi 1993

Reduction of S-N Curves for Ship Structural Details by K. Stambaugh,D. Lesson, F. Lawrence, C-Y. Hou, and G. Banas 1993

Underwater Repair Procedures for Ship Hulls (Fatigue and Ductility ofUnderwater Wet Welds) by K. Grubbs and C. Zanis 1993

Establishment of a Uniform Form-at for Data Reportinq of StructuralMaterial Properties for Reliability Analysis by N. Pussegoda, L. Malik,and A. Dinovitzer 1993

Maintenance of Marine Structures: A State of the Art Summary byS. Hutchinson and R. Bea 1983

Loads and Load Combinations by A. Mansour and A. Thayamballi 1884

Effect of Hiqh Strermth Steels on Stren@h Consdierations of Desicm andConstruction Details of Ships by R. Heyburn and D. Riker 1994

Ship Structure Committee Publications - A Special Bibliography L-