8/13/2019 fatigue failure of welded joints
1/40
Lecture
u si n
atigue ailure Welded oints
8/13/2019 fatigue failure of welded joints
2/40
Lt cture ope undamentals of fatigue failure ofmetals
ffects walding on fatigue atigue design approaches
8/13/2019 fatigue failure of welded joints
3/40ecture 8
Many types of structure experience fluctuatingor repetitive loading ridgl S Axlesor shafts in machinery and vehicles Pressure vessels and piping inydi operation
p
8/13/2019 fatigue failure of welded joints
4/40
Example o t ~ t i n axle
Stress history at point
uf---I-----.,---- ------ \_at;
TIME,
8/13/2019 fatigue failure of welded joints
5/40Lecture
atigue FaHure
Fatigue failure is the formation and growth ofa craGk caused by repeated or fluctuatingloading Continued crack growth may end in suddencollapse or fracture when the remaining areais insufficient to support the load
p
8/13/2019 fatigue failure of welded joints
6/40
F t C -11 .a Jgue.rac : . n ~ r o o n If the stress range is sufficiently high, plastic slip occurs insurface grains After a number of cycles microscopic cracks initiate at theslip regions and at microscopic defects Stage 1 cracks are slow to initiate and grow
Grain BoundaI)urf ceMicro Cracks
p6
8/13/2019 fatigue failure of welded joints
7/40
fngrt le rack rowth Stage 2 cracks are less influenced bymicrostructure They tend to be oriented normalto the maximum tensile stress Each load cycle produces a crack growthincrement The magnitude of the growth incrementdepends on the stress intensity material
propel1ies and environment
Lecture 18
8/13/2019 fatigue failure of welded joints
8/40
Fatigu9Strength: S plot Fatigue strength is commonly represented by a plot of stressrange against cycles to failure or S-N plot However, before S-N data can be used the designer has tohave a y of accounting for the relevant stresses
(
Fatigue life at stress range SI is Nl cycles Se is endurance limitbelowwhich alternating stresses are considered . _.. - _. _. _. non-damagingr, -L : e3E+3 lE 4 3E+4 lE 5 3E+5 lE 6 3E+6No. of Cycles N
8 p 8
8/13/2019 fatigue failure of welded joints
9/40cture 18
h lo ding c n e descri ed in terms of the ratio o maximum to minimum stress
or mean stress and the stress range
8/13/2019 fatigue failure of welded joints
10/40
Nature of The Stress VariationStress Ratio in ax
Fluctuating tension R=O.5
Pulsating tension R=O
enen l ~ \ \ /J) Alternating stress R =-1
kvvv PUlsating- compression R=O
p l0
8/13/2019 fatigue failure of welded joints
11/40
ture of The Stress ~ r i a t k j ; ~
Increasing stress range reduces cyclic life asshown by S N plot Increasing mean stress reduces cyclic life for agiven stress range Cyclic life is practically independent offrequency of loading or the shape of theloading cycle
Lecture 18 p
8/13/2019 fatigue failure of welded joints
12/40
umulative Fatigue Damaqe Variable ampl tude loading is commonly accounted for byMiner's Rule:
(
n+
Miners rules states that- the fatigue damage at any particular stress isproportional to the number of cycles niaccumulated and the cyclic life Ni at that stress- The damago accumulates linearly until failureoccurs
Only approxirr ately accurate Various methods used for counting load cycles inrandom loadings, e.g. rainflow method
p12
8/13/2019 fatigue failure of welded joints
13/40cture 18
t r e s ~ C o n c e 1 t r a t ~ o n :
Changes geometry and stiffness producelocal regions of higher stress termed stressconcEmtration The magnitude of the stress concentrationvarieB with the size of the detail and itssharpness Fatigue initiation is sensitive to local peak
stresBes at stress concentrations
p
8/13/2019 fatigue failure of welded joints
14/40
orrosion atigue~ _ . ~ _ . _ Much fatigue data is based on tests in air Metals may display significantly reduced fatigue strength in other environments
- E.g. ASME Boiler Pressure Vessel Code fatigue designcurves founcl to be non-conservative for steels in hightemperature water- Sea water reduces fatigue strength of welded tubularconnections offshore oil rigs
Termed corrosion fatigue Fatigue data for the specific environmentshould be used.
L _
(
8 p
8/13/2019 fatigue failure of welded joints
15/40
Lecture 18
Hngh Tensile t e e ~ s Under ideal conditions fatigue strengthincreases with yield strength but this is not trueof welded joints e l d l ~ d specimens of high tensile steels andlower strength mild steel display similar S Ncurves. The advantage of high strength steels isreduced when fatigue is a consideration ifdesign stresses are limited by cyclic life
p
8/13/2019 fatigue failure of welded joints
16/40
atigue Susceptibility Susceptibility to fatigue depends therefore onthree basic factors:- repeated or fluctuating loads- the number of loading cycles- stress conce ntr tion
fthese three factors, the one most influencedby designers is the third, through the choiceofdesign detclils More bluntly, this is the one designers mostoften get wrong
P 6
8/13/2019 fatigue failure of welded joints
17/40
Unlike bolted or rivetedattachments welds form anintegral part of a structure Fillet welded bracketsstiffeners etc. prodUcesevere local stressconcentrations due to thesudden change in shape
ource Richards K.G.: Fatigue ofWaided 5IructureoThe Welding Inslitute 1 l69
Lecture 8 P
8/13/2019 fatigue failure of welded joints
18/40
ffe ts elding atigue It is easy to create welded details that produce stressconcentrations simply because of the arrangement of material The hot spots arrowed on the stiffened panel and saddle-supported vessel are potential sites for fatigueinitiation
Stress Concentrations
p
8/13/2019 fatigue failure of welded joints
19/40
V
L l Yr \_{ lNfl ~. -I I
Lecture 18
Groove , ~ J e ~ d s6
~ The fatiguEl strength of 1groove welds transverse to llithe f1uctua ting stress can be g12related to the stress concentration at the edges ;of the welel bead ~ 8
Additional details that 8reduce fatigue strength E 4include misalignment:>notches or excessive Cl 2~reinforcem ent u
12 4 8REINFORCEMENT NGLE
p
8/13/2019 fatigue failure of welded joints
20/40
roov \rH\felds -- ._ ___ _acking strip
left in placeMisalignment
xcessive rootreinforcement
Lack of fusion
, -, l~ ~ ~ ~ ;; :_ -- ----- ~ = - ~ - _ = :0 = = = ~ C _
~ . ~:: ::: : = -- - - ~ = = = r . t =
p20
8/13/2019 fatigue failure of welded joints
21/40
Lecture 18
Fillet welds cause more fatigue problems thangroove welds for two reasons:Their inherent shape produces more severe stressconamtrations
The flexibility they allow in detail design encourages theuse c f gussets brackets and other miscellaneousattachments on load carrying members
P
8/13/2019 fatigue failure of welded joints
22/40
illet eldsEven fillet welds that carry no load reduce fatiguestrength due to tneir effect on the profile and stressconcentration the load carrying part ~
Load carrying fillet welds onc ve filet
Non load carrying weldsJ
18 P 22
8/13/2019 fatigue failure of welded joints
23/40Lecture 18
The location of cracking in load carrying filletwelds depends o the ratio of stress in theweldto the stress the base metal. If the weld is highly stressed cracks initiate atthe root of the weld Making the welds bigger increases fatigue
s t r n ~ l t h until cracking initiates at the weld toes. Beyond this increases weld size do notincreclse fatigue strength
p
8/13/2019 fatigue failure of welded joints
24/40
illet elds
There is little advantage to begained by making fillet weldedattachments parallel to thedirection of stress
yII Attachments made to theedge of a stres ed memberhave even lowE r fatiguestrength than a tachments tothe plate surfac e J
6 p
8/13/2019 fatigue failure of welded joints
25/40cture 8
f f e c ~ of Weld e s i d u ~ Stress
Welds may contain tensile residual stresses upto yiel j strength in magnitude Resid l stresses act as a mean stress andreduci the fatigue strength of the joint Resid l stresses can result in fatigue failuresof wei jed joints even when the loading isentirely compressive
p
8/13/2019 fatigue failure of welded joints
26/40
eld efects The fatigu strength of groove and fillet weldsis governed primarily by their external pr9file
Internal f d defects such as slag inclusions orporosity W thin normal standards ofworkmans lip have little effect on fatiguestrength However i1 butt welds where the reinforcementhas been f emoved the fatigue strength canapproach that of the parent plate. Internal welddefects me y then come into play and reducefatigue life
p
8/13/2019 fatigue failure of welded joints
27/40
\ e ~ d Fatigue ~ m p r o v n n t
Desi ~ n to avoid stress concentration and poorfatig le details on highly loaded members Impr ve weld profile by grinding to blend withsurfe e
Reduce residual stress by heat treatment orothe means
Lecture 8 p
8/13/2019 fatigue failure of welded joints
28/40
8
atigue esign eldment s Three s i l ~ methods for design Nominal stl ess2 Geometric hot spot stress3 Notch stress
Each methl)d estimates the fatigue strengthfrom differe nt levels of detailed informationabout the joint Each method must be used with appropriatedata for fatigue resistanceReference Fatigue c esign of welded joints and components IIW documentXIII-1539-96, Abingtoll Publishing, 1996
p 8
8/13/2019 fatigue failure of welded joints
29/40cture 18
Nominal St.ress Methoc The nominal stress in the member is comparedagains fatigue resistance tabulated for differentstructural details terms of S N curves The nominal stress method is used by severalstandards e.g. AWS 01.1 and eS W59 fordynamically loaded steel structures such asbridges
P
8/13/2019 fatigue failure of welded joints
30/40
Nominal Stress Nominal stress is the maximum stress calculated in thecross section disregarding local stress concentration effects
but including the effects the macrogeometric shape thecomponent e.g. large cut-outs
a b
d ~ ~ e ~~ l t J
P 30
8/13/2019 fatigue failure of welded joints
31/40
F trgl Categori ?s Examples
ture 18
Joint Detailc ~C ~
~
~ ~~ ~
Stress Category
BB ground flush and NDE)C NDE)B,C,D,E dependingon R see tables)C D EF
p
8/13/2019 fatigue failure of welded joints
32/40
sign Stress ange urv s
4-100
2IV0.- -_L , ; ; ; ; ; ; ; ; ; ; ; ; ; ; 100 en
: ; ; ; ; ; ; ~ ~ ~ ~ : = ~ 5 0 CCATGORYD CATEGORY E l
2 O ~
S-lQO 10 2 10Y LI LIFE
66403
15enc g~ 5~ 4 3
2
1 10
p32
8/13/2019 fatigue failure of welded joints
33/40
~ o m e t r i c Stress ethod
The structural geometric stress includes allstress raising effects of the joint geometry butexcludes stress concentrations due to the welditself The stress is compared against S N curves forthe fatigue resistance of the joint detail
Lecture 18 P
8/13/2019 fatigue failure of welded joints
34/40
eometri tr ss E x a m o k ~ sI.. . ~ - - - - - - - --_._ -- --
c d
8
L ~ l L
P 34
8/13/2019 fatigue failure of welded joints
35/40
Lecture
Georruetric tress Methc d .. , > .. . . ,_ 0._ ,,,,,, ._., ~ . . > . .
8/13/2019 fatigue failure of welded joints
36/40
Notch Stress ethod Effective notch stress is the stress at the root of
the notch assuming linear elastic e h v i ~ u r The notch stress is compared with fatigueresistance in terms of a universal S curve forthe materialot h st ss
l : 7concentrations
P 36
8/13/2019 fatigue failure of welded joints
37/40
Notch Stress nnethod
Lecture 18
The ASME B PV Section III effect uses thenotch stress method for fatigue assessment ofwelds nuclear pressure vesselsApplies a stress concentration factor for the weld detail tothe calculated geometric stress in the vessel shellCompares peak stress against a universal S N curve
p
8/13/2019 fatigue failure of welded joints
38/40
Steam enerator ateralRestraint ugs- Attachments to SG shell- Subject to thermal gradientstresses during reactor start-up
and shutdown
x mple atigue ssessment._.... ,. . M ..... , , ~ _ . ,. ._ ........ , _ , ~ , . . , _ . ~ , . , _ . ... , ,. . . = , . . . ~ - - . , ~ - .... , .. A /
~ ~ ~ _ B 3 1
Ddlil ofLalenl _supports t 0 -
P 38
8/13/2019 fatigue failure of welded joints
39/40
Lecture
Steam enerator Load ycle
p
8/13/2019 fatigue failure of welded joints
40/40
SG Lug atigue nalysis _ighly stressed re s