Upload
julio-banks
View
507
Download
4
Embed Size (px)
Citation preview
© 2014 Grzegorz Glinka. All rights reserved. 1
Advances in Fatigue and Fracture Mechanics
June 2nd – 6th, 2014,
Aalto University,
Espoo, Finland
given by
Prof. Darrell Socie
University of Illinois, USA
Prof. Grzegorz Glinka
University of Waterloo, Canada
© 2014 Grzegorz Glinka. All rights reserved. 2
1. Bannantine, J., Corner, Handrock, Fundamentals of Metal Fatigue Analysis, Prentice-Hall,1990.... (good general reference)
2. Dowling, N., Mechanical Behaviour of Materials, Prentice Hall, 2011, 3rd edition(middle chapters are a great overview of most recent approaches to fatigue analysis),
3. Stephens, R.I., Fatemi, A., Stephens, A.A., Fuchs, H.O., Metal Fatigue in Engineering, JohnWiley, 2001.... (good general reference),
4. M. Janssen, J. Zudeima, R.J.H. Wanhill, Fracture Mechanics, VSSD, The Netherlands, 2006(understandable, rigorous, mechanics perspective),
5. Socie, D.F., and Marquis, G.B., Multiaxial Fatigue, Society of Automotive Engineers, Inc.,Warrendale, PA, 2000
5. Haibach, E., Betriebsfestigkeit, VDI Verlag, Dusseldorf, 1989 (in German).
6. Bathias, C., and Pineau, A., Fatigue des Materiaux et des Structures, Hermes, Paris, 2008 (inFrench and English),
7. Radaj, D., Design and Analysis of Fatigue Resistant Structures, Halsted Press, 1990,(Complete, civil and automotive engineering analysis perspective),
8. V.A. Ryakhin and G.N. Moshkarev, Durability and Stability of Welded Structures in Earth MovingMachinery”, Mashinostroenie, Moscow, 1984 (in Russian, cranes and earth moving machinery),
9. A. Chattopadhyay, G. Glinka, M. El-Zein, J. Qian and R. Formas, Stress Analysis and Fatigue ofWelded Structures, Welding in the World, (IIW), vol. 55, No. 7-8, 2011, pp. 2-21.
Bibliography
© 2011 Grzegorz Glinka. All rights reserved. 3
Mechanical Engineer – yesterday and today……
Slide ruler Calculator Computer PC/laptop
1-2 operation/min. 1-10 operation/min. 10? operation/min.
Before yesterday Yesterday Today
© 2008 Grzegorz Glinka. All rights reserved. 4
DAY 1
Contemporary Fatigue AnalysisMethods
(basics concepts and assumptions)
Information Path for Strength and Fatigue Life Analysis
ComponentGeometry
LoadingHistory
Stress-StrainAnalysis
Damage Analysis
Fatigue Life
MaterialProperties
© 2008 Grzegorz Glinka. All rights reserved. 5
Stress Parameters Used in Fatigue Analyses
© 2008 Grzegorz Glinka. All rights reserved. 6
K YS a
a)
Sn
y
dn
0
T
epeak
S
y
dn
0
T
peak
M
b) c)
y
a0
T
Ka2
S
S
apeak
epeak
Sn crack
Sn – net nominal stress; S – gross nominal stresse
peak – local linear-elastic notch-tip stressa
peak – local actual elastic-plastic notch-tip stress
Kt = epeak /Sn – stress concentration factor
K – stress intensity factor
What stress parameter is needed for the FractureMechanics based (da/dN- K) fatigue analysis?
x
a0
T
Stre
ss(x
) 2x
xK
S
S The Stress Intensity Factor K characterizingthe stress field in the crack tip region isneeded!
The K factor can be obtained from:- ready made Handbook solutions (easy to usebut often inadequate in practice)
- from the near crack tip stress (x)distribution or the displacement data obtainedfrom FE analysis of a cracked body (tedious)
- from the weight function by using the FEstress analysis data of un-cracked body(versatile and suitable for FCG analysis)
© 2008 Grzegorz Glinka. All rights reserved. 7
Loads and stresses in a structure
n
Load F
peak
© 2008 Grzegorz Glinka. All rights reserved. 8
The Most Popular Methods forFatigue Life Analysis - outlines
• Stress-Life Method or the S - N approach;uses the nominal or simple engineering stress ‘ S ‘ toquantify fatigue damage
• Strain-Life Method or the - N approach;uses the local notch tip strains and stresses to quantifythe fatigue damage
• Fracture Mechanics or the da/dN - K approach;uses the stress intensity factor to quatify the fatiguecrack growth rate
© 2008 Grzegorz Glinka. All rights reserved. 9
Information path for fatigue life estimation based onthe S-N method
LOADING
F
t
GEOMETRY, Kf
PSO
MATERIAL
0
E
Stress-StrainAnalysis
Damage Analysis
Fatigue Life
MATERIAL
No N
n
eADF
© 2008 Grzegorz Glinka. All rights reserved. 10
a) Structure
Q
H
F
K 5
n
The Similitude Concept states that if the nominal stress histories in the structure and in the testspecimen are the same, then the fatigue response in each case will also be the same and can bedescribed by the generic S-N curve. It is assumed that such an approach accounts also for the stressconcentration, loading sequence effects, manufacturing etc.
K0K1K2K3K4K5St
ress
ampl
itude
,n/2
orhs
/2Number of cycles, N
0
The Similitude Concept in the S-N Method
© 2008 Grzegorz Glinka. All rights reserved. 11
Steps in Fatigue Life Prediction Procedure Based on theS-N Approach
The S – N method
5
e) Standard S-N curves
K0K1K2K3K4K5St
ress
ampl
itude
,n/2
orhs
/2
Number of cycles, N
d) Standard welded joints
c) Section with welded joint
a) Structure
b) Component
VP
R
Q
H
F
Weld
© 2008 Grzegorz Glinka. All rights reserved. 12
5
5
5
5
5
11
1 5
22
2 5
33
3 5
44
4 5
55
5 5
1
1
1
1
1
m
m
m
m
m
DN C
DN C
DN C
DN C
DN C
1 2 3 4 5 ;D D D D D D1
2
3 45
Stre
ss
t
g)
Fatigue damage:h)
Total damage:i)
Fatigue life: N blck=1/Dj)
Stre
ss,
nf)
t
K5
n
Steps in Fatigue Life Prediction Procedure Basedon the S-N Approach (continued)
© 2008 Grzegorz Glinka. All rights reserved. 13
© 2008 Grzegorz Glinka. All rights reserved. 14
The linear hypothesis of Fatigue Damage Accumulation(the Miner rule)
1
11
1 ;m
DN A
31 52 4
1 1 ;1 1 1 1 1R RRfL
NL
NN
Nn
D NN
2
22
1 ;m
DN A
4
44
1 ;m
DN A
51 0;D 42 51 3
5
51
32 41
1 1 1 1 1 ;
1 !!
ii
D
N
D D
N
D
N
D D
NN
D
if D Failure
e
N0N2N1 N3
Stre
ssra
nge,
N4
N5
Cycles
Ni( i)m = A or Ni=A/( i)m
nR
3
33
1 ;m
DN A
The FALSN fatigue life estimation software – Typical input and output data
Weldment
© 2008 Grzegorz Glinka. All rights reserved. 15
The scatter in fatigue:Fatigue S-N curves for assigned probabilityof failure; P-S-N curves
Stre
ssam
plitu
deS a
[N/m
m2 ]
Number of cycles N
400
450
500
105 106 107
(source: S. Nishijima, ref. 39)
S45 Steel temperedat 600o C (W1)
Probability of failureP(%)
P=99%
P=90%
P=50%
P=10%
P=1%
8
648
108
f(N)
f(S)
ma A N
© 2008 Grzegorz Glinka. All rights reserved. 16
N
j < LT
COMPUTING of Tj
SAMPLING A SET OF RANDOM DATA
fK
KfK
fR
RwR
fS
SmaxS
STRUCTURAL COMPONENT
PWL
MATERIAL
No N
S
Rw
LOADING
S
t
Failure probability calculationPf = P[T(X) Tr]:
( )ff
L T TrP
L
Probabilistic fatigue life assessment
© 2008 Grzegorz Glinka. All rights reserved. 17
Over designedUn-satisfactory Mostfrequent
Characteristic regions of cumulativeprobability of the fatigue life distribution
© 2008 Grzegorz Glinka. All rights reserved. 18
LOADING
F
t
GEOMETRY, Kf
PSO
MATERIAL
0
E
Stress-StrainAnalysis
Damage Analysis
Fatigue Life
Information path for fatigue lifeestimation based on the -N method
MATERIAL
2Nf
© 2008 Grzegorz Glinka. All rights reserved. 19
a) Specimen
b) Notched component
peakx
y
z
peak
peak
''2 2
2bf
f f fN N cE
log
(/2
)
log(2Nf)
f
0
l)
j)
The Similitude Concept states that ifthe local notch-tip strain history in thenotch tip and the strain history in thetest specimen are the same, then thefatigue response in the notch tip regionand in the specimen will also be thesame and can be described by thematerial strain-life ( -N) curve.
The Similitude Concept in the – N Method
© 2008 Grzegorz Glinka. All rights reserved. 20
Steps in fatigue life prediction procedure based onthe - N approach
a) Structure
b) Component
c) Section with welded joint
d)
peak
n
peak
hs
nhs
© 2008 Grzegorz Glinka. All rights reserved. 21
Fatigue damage:
1 2 3 41 2 3 4
1 1 1 1; ; ; ;D D D DN N N N
Total damage:
1 2 3 4 ;D D D D D
Fatigue life: N blck=1/D
e)
log
(/2
)
''2 2
2b cf
f f fN NE
e/E
log(2Nf)
f/E
f
02Ne2N
2
: peakNeuberE
' '
' '
'
'
,
,
,
?,?
f f
f f
f
f
fN
p e=
peak
t
1
2
3
4
5
6
7
8
1'
f)
0t
'1
'
n
E K
0
3
2,2'4
5,5'7,7'
68
1,1'
(continued). Steps in fatigue life prediction procedure based on the -N approach© 2008 Grzegorz Glinka. All rights reserved. 22
© 2008 Grzegorz Glinka. All rights reserved. 23
© 2008 Grzegorz Glinka. All rights reserved. 24
© 2008 Grzegorz Glinka. All rights reserved. 25
Three basic sets of input data for the evaluation of the Fatigue CrackInitiation Life and Reliability (the – N approach)
T T
R(T) = 1 - F(T)1f(T)
LOADINGS
t
f(k)
Scaling factor k
LIFE CALCULATION: Ti
i < LY
SAMPLING: k , Kt , f , f' , K’
N
COMPONENT
f(Kt)
SCF Kt
K5
S S
Computing of failure probabilities( )f r
f r
L T TP P T X T
L
MATERIAL
2N 1
K'
fM1
’fe
fM3
K’M
fM2
’fs
2p
2,f
,f
E
2
© 2008 Grzegorz Glinka. All rights reserved. 26
Prob
abili
tyof
failu
re
© 2008 Grzegorz Glinka. All rights reserved. 27
Information path for fatigue life estimation based onthe da/dN- K method
LOADING
F
t
GEOMETRY, Kt
PSO
MATERIAL
0
E
Stress-StrainAnalysis
Damage Analysis
Fatigue Life
MATERIAL
n
KKth
dadN
© 2008 Grzegorz Glinka. All rights reserved. 28
The Similitude Concept states that if the stressintensity K for a crack in the actual componentand in the test specimen are the same, then thefatigue crack growth response in the componentand in the specimen will also be the same andcan be described by the material fatigue crackgrowth curve da/dN - K.
a) Structure
Q
H
F
ab) Weld detail
c) Specimena
P
P
K 10-12
10-11
10-10
10-9
10-8
10-7
10-6
1 10 100
Cra
ckG
row
thR
ate,
m/c
ycle
mMPa,K
The Similitude Concept in the da/dN – K Method
© 2008 Grzegorz Glinka. All rights reserved. 29
© 2008 Grzegorz Glinka. All rights reserved. 30
VP
Q
R
H
F
Weld
a) Structure
b) Component
c) Section with welded joint
S 1
S 2
S3 S4
S 5
Stre
ss,S
t
e)
d)
Steps in the Fatigue Life Prediction Procedure Basedon the da/dN- K Approach
Stress intensity factor, K(indirect method)
Weight function, m(x,y)
, ,A
n
K x y m x y dxdy
KYa
Stress intensity factor, K(direct method)
2I yFE FE
n
K xor
dUK E EGda
KYa
(x, y)
f)
a
g)
01
mi i i
N
f ii
i
a C K N
a a a
N N
Integration of Paris’ equationh)
af
ai
Number of cycles , N
Cra
ckde
pth,
a
Fatigue Life
i)
Steps in Fatigue Life Prediction Procedure Based on theda/dN- K Approach (cont’d)
© 2008 Grzegorz Glinka. All rights reserved. 31
© 2008 Grzegorz Glinka. All rights reserved. 32
© 2008 Grzegorz Glinka. All rights reserved. 33
© 2008 Grzegorz Glinka. All rights reserved. 34
No
j < LYes
CALCULATION of Tj
SAMPLING RANDOM VARIABLES
fa
aa
fC
CC
fS
SmaxS
STRUCTURAL COMPONENT
PWL
MATERIALLOADING
S
t
Failure probability calculationPf = P[T(X) Tr]:
L)TrT(LP f
f
fKth
KthKth
fK
KtK
n
K
da
Kth
dN
fKc
KIcKc
Probabilistic analysis using MC simulation
The FALPR statistical simulation flow chart for the analysis of fatigue crack growth© 2008 Grzegorz Glinka. All rights reserved. 35
Irregular geometrical shape of a real fatigue crack
© 2008 Grzegorz Glinka. All rights reserved. 36
Welded Joint with Load = 4000 lb (ai/ci=0.286)
0
50000
100000
150000
200000
250000
300000
0 0.5 1 1.5 2 2.5
Crack Length 2c (in)
Fatig
uelif
eN
f(C
ycle
)
Exp. (Specimen #11)Exp. (Specimen #12)Exp. (Specimen #14)Exp. (Specimen #15)Exp. (Specimen #16)Exp. (Specimen #17)Exp. (Specimen #18)Pre. Life (Loc. 1) No RsPre. Life (Loc. 1) With Rs
© 2008 Grzegorz Glinka. All rights reserved. 37
© 2008 Grzegorz Glinka. All rights reserved. 38
Global and Local Approaches to StressAnalysis and Fatigue
© 2010 Grzegorz Glinka. All rights reserved. 39
22a
11a
32a=0
33a
31a 13
a= 0
12a= 0
21a=
023
a
11
22
33
0 00 00 0
a
a aij
a
Stress state near the notch tip (on the symmetry line)
1
2
3
© 2010 Grzegorz Glinka. All rights reserved. 40
22a
11a= 032
a
33a
31a 13
a= 0
12a= 0
21a=
023
a
11
22 23
32 33
0 000
a
a a aij
a a
Stress state in the disk at the blade-disk interface
1
2
3
© 2010 Grzegorz Glinka. All rights reserved. 41
1
peak
n
22
11
33
3
2
A D B
C
F
F
A, B
22
22
22
22
33
11
C
D
nnet
p t neak
FA
andK
Stresses concentration in axis-symmetric notched body
© 2010 Grzegorz Glinka. All rights reserved. 42
Stresses concentration in a prismatic notched body
net
pea
n
ntk
FA
andK
A, B, C
22
22
22
33
D
E
11
peak
n
22
11
33
3
2
A D B
C
F
F
1E
© 2010 Grzegorz Glinka. All rights reserved. 43
How to get the nominal stress from the FiniteElement Method stress data?
The smallest optimum elementsize < ¼r !!
r
© 2008 Grzegorz Glinka. All rights reserved. 44
Loads and stresses in a structure
;
;,
, ?
?i
i
ak
n
p
i
e i F F
F F
f f
g g
F
t
Fi
Fi+1
Fi-1
0
n pea
k
© 2008 Grzegorz Glinka. All rights reserved. 45
Loads and StressesThe load, the nominal stress, the local peak stress and the stress concentration factor
, ,
;
;
; ;
;
;
;
1
;
peaktF F
Fpeak i i
F
Fn i i
nnet
n F
Fn
Fpe
et
ak
h h KF
hk F F
k
FA
k F
k A
h F
Axial load – linear elastic analysis
n
y
dn
0
T
peak
Stre
ss
F
F
Analytical, FEM Hndbk
© 2008 Grzegorz Glinka. All rights reserved. 46
Loads and StressesThe load, the nominal stress, the local peak stress and the stress concentration factor
, ,
,
;
;
; ;
btMpeak i pe Mi ak i i
netn
net
n M
netM Mn i i
net
M or kh M K
M cI
k M
ck k MI
Bending load – linear elastic analysis
n
y
dn
0
T
peak
Stre
ssM
b) M
© 2008 Grzegorz Glinka. All rights reserved. 47
peak eakt t
n
poK rS
K
Kt – stress concentration factor(net or gross, net Kt gross Kt !! )
peak – stress at the notch tip
n - net nominal stress
S - gross nominal stress
Stress Concentration Factors in Fatigue AnalysisThe nominal stress and the stress concentration factor in simple load/geometry configurations
nnet gross
P Por SA A
grossnetn
net gross
M cM cor S
I I
Simple axial load
Pure bending load
n
y
dn
0
T
r
peak
Stre
ss
S
S
Tension
n
y
dn
0
T
r
peak
Stre
ss
MS
SM
Bending
net Kt gross Kt
© 2008 Grzegorz Glinka. All rights reserved. 48
Stress concentration factors for notched machine components
(B.J. Hamrock et. al., ref.(26)
d b
h
1.0
1.4
1.8
1.2
1.6
2.0
2.2
2.4
2.6
2.8
3.0St
ress
conc
entr
atio
nfa
ctor
Kt=
peak
/n
Diameter-to-width ratio d/b
0 0.1 0.2 0.3 0.4 0.5 0.6
peak
P P
n =P/[(b-d)h]
© 2008 Grzegorz Glinka. All rights reserved. 49
Stress concentration factors for notched machine components
(B.J. Hamrock et. al.
1.0
1.4
1.8
1.2
1.6
2.0
2.2
2.4
2.6
2.8
3.0
0 0.05 0.10 0.15 0.20 0.25 0.30
H/h=6H/h=2H/h=1.2H/h=1.05H/h=1.01
Stre
ssco
ncen
trat
ion
fact
orK
t=pe
ak/
n
Radius-to-height ratio r/h
peak
r
Mh
b
MH
26
nMc MI bh
© 2008 Grzegorz Glinka. All rights reserved. 50
Various stress distributions in a T-butt weldment with transverse fillet welds;
r
t
t1
ED
BC
A
pea
k
n
hs
FP
M
C
• Normal stress distribution in the weld throat plane (A),• Through the thickness normal stress distribution in the weld toe plane (B),• Through the thickness normal stress distribution away from the weld (C),• Normal stress distribution along the surface of the plate (D),• Normal stress distribution along the surface of the weld (E),• Linearized normal stress distribution in the weld toe plane (F).
Stress concentration & stress distributions in weldments
© 2010 Grzegorz Glinka. All rights reserved. 51
0.65
1 exp 0.92 11 2
2.8 21 exp 0.452
tent
Wh hK
W rWth
: 2 0.6 pwhere W t h h
Range of application - reasonably designed weldments, (K.Iida and T. Uemura, ref. 14)
Stress concentration factor for a butt weldmentunder axial loading
g=
h
r
t
l = hp
PP
© 2010 Grzegorz Glinka. All rights reserved. 52
Stress concentration factor for a T-buttweldment under tension load; (non-load carrying fillet weld)
0.65
1 exp 0.92 11
2.8 21 exp 0.452
tt
Wh hK
W rWth
: 2 0.3 2p pwhere W t h t h
Validated for : 0.02 r/t 0.16 and30o 60o
, source [14]
t
r
t1= tp hp
h
PP
y
x
Cyclic Loads and Cyclic Stress Patterns(histories) in Engineering Objects
© 2008 Grzegorz Glinka. All rights reserved. 53
Loads and StressesThe load, the nominal stress, the local peak stress and the stress concentration factor
,
,
, ;
;
;i
bti
t
F Mn i i
tpeak i
F Mpea
i
F M
k
i
ii
M
M
K
k F k
K
h
M
k F kor
hF
Simultaneous axial and bending load
n
y
dn
0
T
peak
Stre
ss
M
b) M
F
F
;
,
,
,
;
;
bt
tt
F M
F M
K
h
k
K
h
k From simple analytical stress analysis
From stress concentration handbooks
From detail FEM analysis
© 2008 Grzegorz Glinka. All rights reserved. 54
3
3
32;
; ; ;4 2 64
b b
b
eakn pM c M
SI d
W d dM L l c INote! In the case of smooth components,
such as the railway axle, the nominal stress and the local peak stress are the same!
b)
d N.A.
,min 3
32n
bMd
,min 3
32n
bMd
1
2
3
n,max
n,min
timeStre
ss n,a
1 cyclec)
1
2
3
AB
y
x
Ll
Moment Mb
a)
RARB
W/2W/2 W/2
W
The load W and the nominal stress n in an railway axle
© 2008 Grzegorz Glinka. All rights reserved. 55
22
A B C
t
t
23
0
0
A
B
C
23
2
20
A B C
t
t
23
0
0
22
A
B
C
23
22
0
Non-proportional loading path Proportional loading path
F2R
t
x2
x3
F
T
T
22
33
22
33
x2
x3
23
Fluctuations and complexity of the stress state at thenotch tip
© 2008 Grzegorz Glinka. All rights reserved. 56
How to establish the nominal stress history?a) The analytical or FE analysis should be carried out for one characteristic load magnitude, i.e.P=1, Mb =1, T=1 in order to establish the proportionality factors, kP, kM, and kT such that:
;;P M Tn n nP M Tbk P k M k T
b) The peak and valleys of the nominal stress history n,,i are determined by scaling the peak andvalleys load history Pi, Mb,I and Ti by appropriate proportionality factors kP, kM, and kT such that:
, , ,;,P M Tn i n i n i iiP M Tb ik P k M k T
c) In the case of proportional loading the normal peak and valley stresses can be added and theresultant nominal normal stress history can be established. Because all load modes in proportionalloading have the same number of simultaneous reversals the resultant history has also the samenumber of resultant reversals as any of the single mode stress history.
;,, i Mi Pn b ik P k M
d) In the case of non-proportional loading the normal stress histories (and separately the shearstresses) have to be added as time dependent processes. Because each individual stress historyhas different number of reversals the number of reversals in the resultant stress history can beestablished after the final superposition of all histories.
ii in P M bt tt k P k M
© 2008 Grzegorz Glinka. All rights reserved. 57
Two proportional modes of loading
0
Mode a
Stre
ssn,
b
0
Stre
ssn,
a
0
Mode b
Stre
ssn
Resultant stress: n= n,a+ n,b
0
Mode a
Stre
ssn,
b
0
Stre
ssn,
a
0
Mode b
Stre
ssn
Resultant stress:n(ti)= n,a(ti)+ n,b(ti)
time
time
Two non-proportional modes of loading
Superposition of nominal stress histories induced by twoindependent loading modes
© 2008 Grzegorz Glinka. All rights reserved. 58
Wind load and stress fluctuations in a windturbine blade
Note! One reversal of the wind speed results in several stress reversals
Wind speed fluctuations + Blade vibrations Stress fluctuations
Source [43]
In-plane bending Out of plane bending
Time [s]600 605 610 615 620
Time [s]600 605 610 615 620
Win
dsp
eed
[m/s
]Lo
ad-la
gst
ress
[MPa
]
- 5
5
0
8
10
12
14
- 400
10
12
14
- 50
- 30
- 28
Win
dsp
eed
[m/s
]Lo
ad-la
gst
ress
[MPa
]
© 2008 Grzegorz Glinka. All rights reserved. 59
a) Ground loads on the wings, b) Distribution of the wing bending moment induced by the groundload, c) Stress in the lower wing skin induced by the ground and flight loads
Characteristic load/stress history in the aircraft wing skin
time
Stre
ss
0
Source [9]
0
a)
LandingTaxiing
Flying
b)
c)
© 2008 Grzegorz Glinka. All rights reserved. 60
Loads and stresses in a structure
;
;,
, ?
?i
i
ak
n
p
i
e i F F
F F
f f
g g
F
t
Fi
Fi+1
Fi-1
0
n pea
k
© 2008 Grzegorz Glinka. All rights reserved. 61
Loading and stress histories
Time
0
Stre
sspe
ak
Time0
Load
F
Time
Fi-1
Fi
Fi+1
Fmax
0
Stre
ssn
n,i-1
n,i
n,i+1
n,max
Load history, F Nominal stress history, n Notch tip stress history, peak
0i-1
i
i+1
+1
-1
,n i
max
,
,
,
,
,ma
,
,max,
,
x
1 1
F i
t t F i
i
peak i
peak i
p
n i
n i
n in
eakni
n i
k FK K k F
FF
Characteristic non-dimensional load/stress history
peak,
i
peak,i-1
peak,i+1
peak,max
© 2008 Grzegorz Glinka. All rights reserved. 62
How to get the nominal stress n from theFinite Element method stress data?
Notched shaft under axial, bending and torsion load
a) Run each load caseseparately for an unitload
b) Linearize the FE stressfield for each load case
x3
F
r
D
t
x2F
T
T MMd
Discrete cross section stress distributionobtained from the FE analysis
d 22
0
nx3
peak
© 2008 Grzegorz Glinka. All rights reserved. 63
© 2010 Grzegorz Glinka. All rights reserved. 64
0 0
13 2 2
1 6 62;
112
n
i ib t tn
t y y dy y y dy y y yc M
tI t t
0
2
2
0
1 1
6
6
b tn
n
m
n
i
tn
ii
n
yy dy
t
y y
ydy
t
y y y
tt
0 0
1
1;
1 1
n
im t tn
y dy y dy y yP
t t t t
Determination ofnominal stressesfrom discrete FE databy the linearizationmethod
x
y
yi
(yi)
yiyn
(yn)
(y1)
yn
t1
peakn
t
How to get the resultant stress distribution from theFinite Element stress data? (Notched shaft under axial, bending load)
x3
P22
33
r
D
t
x2
P 23
MM
dd 22
0
nbx3
Bending
d
x3
22
0
nm
Axiald 220
peak
Resultant
x3
(x3)
© 2008 Grzegorz Glinka. All rights reserved. 65
Cyclic nominal stress and corresponding fluctuating stress distribution
Stre
ssn
time
n, max
n, 0
n, min
x3
d
220
Resultant
22(x3, n,max)22(x3, n,0)22(x3, n,min)
© 2008 Grzegorz Glinka. All rights reserved. 66
© 2010 Grzegorz Glinka. All rights reserved. 67
• Multiaxial state ofstress at weld toe
• One shear and twonormal stresses
• Due to stressconcentration, xx isthe largest component– Predominantly responsible
for fatigue damage
zz
xx
xx
zz
zxxz
The stress state at the weld toe
Determination of the nominal, n, and the hot spotstress , hs, from 3D-FE stress analysis data
a) Stress distribution in the critical cross section near the cover plate ending and the nominal or thehot spot stress n (independent of length L ) and hs (independent of length L),b) Stress distribution in the critical plane near the ending of a vertical attachment (gusset) and thenominal or the hot spot stress n (dependent on length L ) or hs (independent of length L)
L
L t
m b thssh
n
st
h
x y dxdyP
t L
x y d x y ydy
t
t
t
y
L
/2 0
/2
00
2
,
6 0,0,
- depends on L and is constant along the weld toe line
Independent of L but it changesalong the weld toe line
y
x
PP
(x,y)
a)
L
tpeak
hs
y
x
PP(x,y)
b)
L
t
© 2008 Grzegorz Glinka. All rights reserved. 68
The Nominal Stress n versus the local Hot Spot Stress hs
y
PP
(x,y)
t
x
A
B
xy
hs,B
hs,A
L
m b m bhs A hsh A hs Bs A hs hs BB, ,, , , ,; ;
n A
n B
Pt L,
,
;
??;
© 2008 Grzegorz Glinka. All rights reserved. 69
Example:Preparation and Analysis of Representative
Stress/Load History:
The Rainflow Cycle Counting Procedure
© 2008 Grzegorz Glinka. All rights reserved. 70
Stress/Load Analysis - Cycle Counting Procedureand Presentation of Results
The measured stress, strain, or load history is given usually in the form of a time series, i.e. a sequence ofdiscrete values of the quantity measured in equal time intervals. When plotted in the stress-time space thediscrete point values can be connected resulting in a continuously changing signal. However, the time effecton the fatigue performance of metals (except aggressive environments) is negligible in most cases.Therefore the excursions of the signal, represented by amplitudes or ranges, are the most importantquantities in fatigue analyses. Subsequently, the knowledge of the reversal point values, denoted with largediamond symbols in the next Figure, is sufficient for fatigue life calculations. For that reason the intermediatevalues between subsequent reversal points can be deleted before any further analysis of the loading/stresssignal is carried out. An example of a signal represented by the reversal points only is shown in slide no. 141.The fatigue damage analysis requires decomposing the signal into elementary events called ‘cycles’.Definition of a loading/stress cycle is easy and unique in the case of a constant amplitude signal as that oneshown in the figures. A stress/loading cycle, as marked with the thick line, is defined as an excursion startingat one point and ending at the next subsequent point having exactly the same magnitude and the same signof the second derivative. The maximum, minimum, amplitude or range and mean stress values characterisethe cycle.
Unfortunately, the cycle definition is not simple in the case of a variable amplitude signal. The only non-dubious quantity, which can easily be defined, is a reversal, example of which is marked with the thick line inthe Figures below.
max min
max min
max min
2
2
a
m s
stress range
stress amplitude
mean stres
max
min
timeStre
ssa
1 cycle
0
m
© 2008 Grzegorz Glinka. All rights reserved. 71
Removing material from a clay mine in Tennessee
© 2008 Grzegorz Glinka. All rights reserved. 72
© 2008 Grzegorz Glinka. All rights reserved. 73
Bending Moment Time Series
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
20
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81
Load point No.
Ben
ding
mom
ent
[10
kNm
]
Bending Moment measurements obtained at constant time intervals
© 2008 Grzegorz Glinka. All rights reserved. 74
Bending Moment History - Peaks and Valleys
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
20
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Load point No.
Ben
ding
mom
entv
alue
[10
kNm
]
Bending Moment signal represented by the reversal point values
© 2008 Grzegorz Glinka. All rights reserved. 75
Constant and Variable Amplitude Stress Histories;Definition of the Stress Cycle & Stress Reversal
max min
max min
max min
min
max
;
2 2
;2
a
m
RStre
ss
Time0
Variable amplitude stress history
Onereversal
b)
0
One cycle
mean
max
min
Stre
ss
Time0
Constant amplitude stress historya)
© 2008 Grzegorz Glinka. All rights reserved. 76
Stress Reversals and Stress Cycles in a VariableAmplitude Stress History
The reversal is simply an excursion between two-consecutive reversalpoints, i.e. an excursion between subsequent peak and valley or valleyand peak.
In recent years the rainflow cycle counting method has been acceptedworld-wide as the most appropriate for extracting stress/load cycles forfatigue analyses. The rainflow cycle is defined as a stress excursion,which when applied to a deformable material, will generate a closedstress-strain hysteresis loop. It is believed that the surface area of thestress-strain hysteresis loop represents the amount of damage inducedby given cycle. An example of a short stress history and its rainflowcounted cycles content is shown in the following Figure.
© 2008 Grzegorz Glinka. All rights reserved. 77
Stre
ss
Time
Stress history Rainflow counted cycles
i-1
i-2
i+1
i
i+2
Stress History and the “Rainflow” Counted Cycles
1 1i i i iABS ABS
A rainflow counted cycle is identified when any two adjacent reversals in thestress history satisfy the following relation:
© 2008 Grzegorz Glinka. All rights reserved. 78
A rainflow counted cycle is identified when any two adjacent reversals inthee stress history satisfy the following relation:
1 1i i i iABS ABSThe stress amplitude of such a cycle is:
1
2i i
a
ABS
The stress range of such a cycle is:
1i iABS
The mean stress of such a cycle is:
1
2i i
m
The Mathematics of the Cycle Rainflow Counting Methodfor Fatigue Analysis of Fluctuating Stress/Load Histories
© 2008 Grzegorz Glinka. All rights reserved. 79
The rainflow cycle counting procedure - example
Determine stress ranges, Si, and corresponding mean stresses, Smi for the stress historygiven below. Use the ‘rainflow’ counting procedure.
Si= 0, 4, 1, 3, 2, 6, -2, 5, 1, 4, 2, 3, -3, 1, -2 (units: MPa 102)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1
2
3
4
5
6
0
-1
-2
-3
Stre
ssS i
(MPa
102 )
Reversing point number, i
© 2008 Grzegorz Glinka. All rights reserved. 80
The ASTM rainflow counting procedure
1. Find the reversing point with highest absolute stress magnitude,
2. The part of the stress history before the maximum absolute attach to the end ofthe history,
3. Perform the rainflow counting on the re-arranged stress history, i.e. frommaximum to maximum
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 2 3 4 5
Original stress history
Absolute maximum !
© 2008 Grzegorz Glinka. All rights reserved. 81
The ASTM modification of the Stress History
1 2 3 4 5 6 7 8 9 11 12 13 14 150
10
123456
-1-2-3
The modifiedstress history
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 2 3 4 5
Absolute maximum
The original stress history
© 2008 Grzegorz Glinka. All rights reserved. 82
The Modified Stress History according to the ASTM
1 2 3 4 5 6 6 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
Stre
ss(M
Pa)x
102
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
starting point
Start counting from point No. 2 !
© 2008 Grzegorz Glinka. All rights reserved. 83
7676 7 6 ,6 7
3 23 2 1; 2.5;2 2m
5454 5 4 ,4 5
1 41 4 3; 2.5;2 2m
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
Start counting from the point No. 2 !!
© 2008 Grzegorz Glinka. All rights reserved. 84
2 32 3 2 3 ,6 7
2 52 (5) 7; 1.5;2 2m
9 109 10 9 10 ,9 10
1 21 ( 2) 3; 0.5;2 2m
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
© 2008 Grzegorz Glinka. All rights reserved. 85
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
13 1413 14 13 14 ,13 14
3 23 2 1; 2.5;2 2m
11 1211 12 11 12 ,11 12
4 14 1 3; 2.5;2 2m
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
© 2008 Grzegorz Glinka. All rights reserved. 86
1 81 8 1 8 ,1 8
6 36 ( 3) 9; 1.5;2 2m
1 2 3 4 5 6 7 8 9 11 12 13 14 15
-3
-2
-1
0
1
2
3
4
5
6
Cycles counted –ASTM method1. 6- 7 =1; m,6- 7 = 2.5;2. 4- 5 =3; m,4- 5 = 2.5;3. 13- 14=1; m,10- 11= 2.5;4. 11- 12=3; m,11- 12= 2.5;5. 2- 3 =7; m,2- 3 = 1.5;6. 9- 10 =3; m,9- 10 =-0.5;7. 1- 8 =9; m,1- 8 = 1.5;
© 2008 Grzegorz Glinka. All rights reserved. 87
Extracted rainflow cycles, m
Total number of cycles, N=854
m -32 -22 -13 -3.2 6.44 16.1 25.7 35.3 45 54 64.1 73.7 83.3 92.9 103 112 122 131 141 151298.8 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1283.9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0268.9 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1254 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1239 0 0 0 0 0 0 0 1 2 2 0 0 0 0 0 0 0 0 0 0 5
224.1 0 0 0 0 0 0 0 0 2 2 1 0 0 0 0 0 0 0 0 0 5209.2 0 0 0 0 0 0 0 3 4 5 2 0 0 0 0 0 0 0 0 0 14194.2 0 0 0 0 0 1 0 1 7 2 0 0 0 0 0 0 0 0 0 0 11179.3 0 0 0 0 0 0 1 0 4 4 0 0 0 0 0 0 0 0 0 0 9164.3 0 0 0 0 1 0 0 0 3 1 0 0 0 0 0 0 0 0 0 0 5149.4 0 0 0 0 0 0 1 0 0 0 0 2 1 0 0 0 0 0 0 0 4134.5 0 0 0 0 0 0 0 0 0 0 0 4 1 1 0 0 0 0 0 0 6119.5 0 1 1 0 0 0 0 0 0 0 3 1 5 1 2 0 0 0 0 0 14104.6 0 0 1 2 1 0 0 0 0 2 4 3 7 3 2 1 2 1 0 0 2989.64 0 1 2 3 7 2 0 0 0 1 2 8 10 7 5 6 2 1 0 0 5774.7 1 1 3 4 3 5 0 1 2 2 10 18 23 20 17 11 4 1 0 0 126
59.76 2 1 5 7 4 1 4 5 1 2 11 20 34 31 31 28 9 7 1 1 20544.82 1 6 9 7 9 7 10 3 3 8 15 37 49 64 62 41 16 11 2 1 36129.88 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 014.94 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
854
Mean stress, m
Stre
ssra
nge,
© 2008 Grzegorz Glinka. All rights reserved. 88
Extracted rainflow cycles, m
© 2008 Grzegorz Glinka. All rights reserved. 89
b) the stress rangefrequency distributiondiagram
Number cycles N
Stre
ssra
nge
jj=7j=6
nj=6 - number of cycles j=6in the class j=6
Total number of cycles in the entire history, NT
max
Rel
ativ
est
ress
rang
ej/
max
Relative number of cycles N/NT
0
j=7j=6j=4321
nj=6/NT
max/ max1
j=4/ max
0.5 1.0
0.5
0
j=4321
a) The stress rangeexceedance diagram(stress spectrum)
© 2008 Grzegorz Glinka. All rights reserved. 90
Day 1
The End
© 2008 Grzegorz Glinka. All rights reserved. 91