Oct 19, 2015
FATIGUE in LWRs
International Conference on Plants Materials Degradation
November 18-20 , 2008
J. MENDEZ (LMPM-ENSMA) J.M. STEPHAN (EDF)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers2
LMPM
PlanPlan
Solicitation and Materials
Codes and Rules
Questions to solve
New Knowledge in Fatigue Behavior of austenitic stainless steel :
High cycle fatigue and fatigue limit
Interaction between Low cycle and High cycle fatigue
Environmental effects
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers3
LMPM
FatigueFatigue
Fatigue is defined as a term which "applies to changes in properties which can occur in
a metallic material due to repeated application of stresses or strains, although usually this term applies specially to those changes which lead to cracking or failure"
(General Principles for Fatigue testing of Metals, 1964)
Fatigue is an ageing process !
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers4
LMPM
FatigueFatigue
Fatigue is an important cause of incidents in NPPs :
Vibrations, Thermal stratification, Vortex in dead legs, Thermalmixing
Fatigue is cited in life extension of NPPs :
Residual life
Upper probability of occurrence due to the accumulation of repeated loadings
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers5
LMPM
Origin of Fatigue in NPPsOrigin of Fatigue in NPPs
Mechanical loadings :
Vibrations (70%)
Thermal loadings
Initially, fatigue incidents were due to large thermal loadings (Thermal Shocks, Stratification)
Nowadays, thermal fatigue problems concern small random thermal cycles due to mixing or vortex
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers6
LMPM
Solicitations and MaterialsSolicitations and Materials Mechanical Solicitations
Vibrations
Most frequent fatigue type occurrence : ~70%
High cycle fatigue failure
Mechanical or flow induced vibration loads
Short time, during plant startup or soon thereafter, or after aging (wear, clearance) Design configurations, modifications, BWR power uprates
Locations
Small bore pipes
Tube bundle (Alliage 600) Main steam systems nozzles for measurements
Valves steam : when not tight close or opened : Valve relief "chattering"
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers7
LMPM
Solicitations and MaterialsSolicitations and Materials Thermal Solicitations
Thermal shocks following modification of plant condition
Not frequent but with high amplitude of temperature (270C) Not localized (hitting the whole component) Vessel, pipes
CVCS nozzle, SIS nozzle, Surge line (AISI 304, CF3, CF8)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers8
LMPM
Solicitations and MaterialsSolicitations and Materials Thermal Solicitations
Thermal Stratification
Inherent under some operating conditions (low flow rates) Possible high stress ranges during flow rate variations Dependence on
bearing modes
Vessel, pipes : Surge line (AISI 304), Feedwater Flow Control Systems (A42, A48, 16MND5)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers9
LMPM
Solicitations and MaterialsSolicitations and Materials
Conclusion on large thermal loads :
Possible measurements on sites (outer wall)Now covered by existing rules and codes
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers10
LMPM
Solicitations and MaterialsSolicitations and Materials Thermal Solicitations
Vortex
Only damaging, if associated with abnormal conditions (valve leakage) Except in the case of non-insulated small diameter pipe
Pipes : "Dead Legs" : SIS, RHR (AISI 304), small pipes linked to Primary Circuit
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers11
LMPM
Solicitations and MaterialsSolicitations and Materials Thermal Solicitations
Turbulent mixing
In general, localizations where thermal fluctuations can take place (piping mixing zones, pumps..)
Inherent under functioning conditions of some systems (pipe junctions, pumps) Possible large damages
Pipes : RHR (AISI 304, 316 ), CVCS (AISI 304, CF8, CF3 ), Pumps (AISI 304, 316, AISI 410)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers12
LMPM
Solicitations and MaterialsSolicitations and Materials
Conclusion on high cycle thermal loads :
No direct measurements on sites due to high frequencies involved
Large non linear stresses gradient
Complexity : links between thermal-hydraulics, mechanics (initiation and fracture), materials and plant operations
Problem of multi-cracks
Combination with low cycle fatigue : damage accumulation
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers13
LMPM
Codes and RulesCodes and RulesASME Code Section III and VIII :
Recent proposed developments :
Revision of fatigue strength curves and margins
Explicit inclusion of environmental effects
Possibility to apply an "initiation + crack growth approach"
can have wide implications in industry
KTA = ASME
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers14
LMPM
Codes and RulesCodes and RulesRCCM :
Derived from ASME but with modifications :
Ke optimisation for stainless steels
Ke for mechanical loads and possible Keth for thermal loads
K for mixing zones
Crack like defects fatigue analysis method
JSME :
Some dedicated standard for low and high cycles fatigue
WWER PNAE Code
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers15
LMPM
Fatigue : Main needsFatigue : Main needs
Definition of screening values with simple parameters
Guidelines to account for complex loadings like random loadings (associated or not with low cycle loads)
Improvements in propagation threshold
Material aspects : Improvements in fatigue curves to take into account surface finishes, mean stresses, welds, residual stresses, environment
Probabilistic aspects in fatigue : random loads
Interaction with variations in metal conditions (ageing, irradiation)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers16
LMPM
Research ProgramsResearch ProgramsLast decades :
Low cycles fatigue (Thermal shocks, Stratification) in pipes, nozzles, pumps
Now :
Major programs devoted to thermal cycling in mixing areas and vortex in dead legs
Thermal-hydraulics
High cycles mechanical testing
Material studies : T, Surface finishes, mean stres ses, environment
Research at the micro and meso level : Dislocations, Aggregates
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers17
LMPM
Questions to solve : Thermal HydraulicsQuestions to solve : Thermal Hydraulics
Experimental :
Quality of on site measurements (external wall) or mock-up measurements (fine measurements on the surface)
Transposition methodologies
Numerical :
Quality of numerical simulations :
R.A.N.S. simulation (mean thermal-hydraulic flows) Large Eddy simulations (LES) (temporal description)
Evaluation of heat exchange coefficients
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers18
LMPM
Questions to solve : MaterialsQuestions to solve : Materials
Material behaviour (plasticity ; strain-stress curves) Initiation Multi axial loading
Initiation - Treatment of surface finishes : roughness and strain hardening (pre loading): How to determine their levels of influence on fatigue
Initiation - Treatment of mean stress effect (in case of large mean stress) Initiation - Damage accumulation with variable amplitude loading
Initiation - Environmental effects
Transposition to on site functional situation
Validation of Fatigue initiation criteria
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers19
LMPM
Questions to solve : MaterialsQuestions to solve : Materials
Propagation :
Propagation Threshold
Short cracks
Overload - Variable amplitude loadings
Multi axial loadings
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers20
LMPM
New knowledge acquired in the recent years on some aspects of the fatigue behavior of austenitic stainless
steels (304L, 316L)New open questions
New knowledge acquired in the recent years on some aspects of the fatigue behavior of austenitic stainless
steels (304L, 316L)New open questions
High cycle fatigue and fatigue limit
effects of surface finish
effects of mean stress
Interaction between LC and HC Fatigue
Environmental effects
Crack initiation and crack growth processes Illustrations from 304L or 316L fatigue data
ENSMA studies in the frame of the PhD of S. Petijean (2003) (EDF, AREVA N.P.)
Y. Lehericy (2007) (AREVA N.P.)L. De Baglion (AREVA N.P.)
examples about :
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers21
LMPM
Even at very low cyclic strain amplitudes in the HCF range, near the fatigue limit, 304L exhibits a significant plasticity at the macroscopic level
J. C. Le Roux, 2004
Main characteristics of 304L austenitic stainless steel behaviour with regard to fatigue behaviour
a high ductility leading to fatigue limit (195-200 MPa) close to the conventional yield stress
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers22
LMPM
INFLUENCE OF SURFACE FINISH PARAMETERS ON THE HIGH CYCLE FATIGUE BEHAVIOUR OF A 304L AUSTENITIC STAINLESS STEEL
INFLUENCE OF SURFACE FINISH PARAMETERS ON THE HIGH CYCLE FATIGUE BEHAVIOUR OF A 304L AUSTENITIC STAINLESS STEEL
Machining or surface operations lead to microstructural and mechanical modifications:
What are the most significant factors that have to be taken into account? :
Roughness : relevant parameters?Residual stresses: relaxation?
Surface layer microstructure modification: delays or accelerates crack initiation and growth?
Coupled effects: How to identify the predominating factors?
The study of several surface preparation conditions permitted to establish the role of surface finish and identify the relative effect of each factor.
(S. Petitjean, ENSMA 2003, EDF and AREVA (Framatome) collaboration)
What are the effects of these modifications?
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers23
LMPM
Machining conditions of some selected surface preparations(polishing, turning and grinding)
Machining conditions of some selected surface preparations(polishing, turning and grinding)
And two polished sample conditions :after fine turning (T5), a mechanical polishing using abrasive papers of grades 320, 500, 1000, 2400 and 4000 followed by diamond sprays of 3 m and 1 m is performed.
Polished P1: only the surface irregularities are suppressed; the hardness gradient is preservedPolished P2: :the hardened layer is also taken off.
Samples Tool radius (mm)Feed rate (mm/rd)Speed (rd/mn) Depth of cut (mm)Turned T1 0,4 0,2 1600 0,2Turned T2 0,8 0,2 1600 0,2Turned T3 0,8 0,5 1120 0,2Turned T4 0,8 0,7 560 0,2Turned T5 0.8 0.7 560 0.2
SandBlasted Silica particles pres. 4 barsGround G1 - Manual 800 Grindstone tangential
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers24
LMPM
Typical roughness profiles of some surface preparationsTypical roughness profiles of some surface preparations
roughness of sample Turned T2
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4length (mm)
r
o
u
g
h
n
e
s
s
(
m
)
roughness of sample Polished P1
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4length (mm)
r
o
u
g
h
n
e
s
s
(
m
)
roughness of sample Turned T3
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4
length (mm)
r
o
u
g
h
n
e
s
s
(
m
)
roughness of sample Ground G1
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4length (mm)
r
o
u
g
h
n
e
s
s
(
m
)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers25
LMPM
Typical appearence of the surface of some 304 L cylindrical fatigue specimens
Typical appearence of the surface of some 304 L cylindrical fatigue specimens
2 mm 2 mm
Turning T3 sandblasting
Coarse grinding polishing
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers26
LMPM
Turned T3rugosit de l'tat tourn 1b
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4
longueur (mm)r
u
g
o
s
i
t
(
m
)
Echantillon tourn T3
longueur (mm)r
u
g
o
s
i
t
(
m
)
With regard to fatigue, a 3D characterization of roughness is required
Sandblasted
- craters distributed on the specimen surface - mean dimensions: 20 m in depth, 100 m in diameter
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4length (mm)
r
o
u
g
h
n
e
s
s
(
m
)
rugosit de l tat sabl
longueur (mm)
r
u
g
o
s
i
t
(
m
)
roughness of sample Ground G1
-32
-22
-12
-2
8
18
28
0 0,5 1 1,5 2 2,5 3 3,5 4
rugosit de ltat meul cylindrique
longueur (mm)
R
u
g
o
s
i
t
(
m
)
2 mm
sets of facets Lmean = 0.5 / 0.6 mm.
crossed by straight grooves
Ground
samples
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers27
LMPM
Associated gradient of hardness in the near surface layer for seven selected sample preparations
Evolution of microhardness Vickers HV25
180
200
220
240
260
280
300
320
340
0 100 200 300 400 500 600depth (m)
m
i
c
r
o
h
a
r
d
n
e
s
s
H
V
2
5
Turned T1Turned T2Turned T3Turned T4Polished P1Polished P2Ground G1
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers28
LMPM
Surfacemicrostructuremodifications:recristallisation, mechanical twinning, high density of
dislocations, martensitic transformations
600 nm
colle
fine recristallisation, dislocations distorded twins and dislocations twins and dislocations(turned, sandblast) (severe turning) (turned, sandblast, polished)
phase transformations ( )
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers29
LMPM
Residual stress profile
Profil des contraintes rsiduelles en fonction de la profondeur
-400
-300
-200
-100
0
100
200
300
400
0 20 40 60 80 100 120Profondeur (m)
C
o
n
t
r
a
i
n
t
e
s
r
s
i
d
u
e
l
l
e
s
(
M
P
a
)
Tourn 1b
Evolution des contraintes rsiduellesen fonction de la profondeur
Tourn T3
- Specimens turned and ground: tensile residual stresses at the specimen surface
- Specimens polished or sandblasted compressive residual stresses
Example of residual stresses profile as a function of depth ( condition Turned T3)
Tensile residual stresses over the first 50 mCompressive below
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers30
LMPM
Condition Ra Rt HV25at 30 m
depth
Depth of the hardened layer
(m)11111111 ((((Pa) 22 22 22 22 (MPa)
phase
(m) (m) transform.?T1 1,6 7,2 323 250 235 300 -T2 1.94 8.85 316 200 220 260 -T3 9.6 38.1 337 250 395 385 -T4 - 78.3 345 250250 375 630 Yes
Polish.P1 0,34 1,44 318 150 -190 -185 YesPolish.P2 0,34 1,44 212 30 -240 -240 YesGround 6,3 36,4 345 200 320 480 -
S.Blasted 3.65 19,85 343 150 -770 -730 Yes
Summary of the surface characterisations of the selected samples by turning, grunding, blasting and polishing
Several combinations of roughness, residual stresses, hardeness and microstructures
Used to identify the relative role of each factor on the fatigue properties
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers31
LMPM
- Fatigue tests under constant load amplitudes control
* S-N curves in air at T = 25C under a constant load ratio (R = 0.05) or at constant mean (mean = 0, 60, 125, 195 MPa)
Seven surface preparations investigated:
3 Turned, 1 sandblasted, 1 severe ground, 2 polished
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers32
LMPM
S-N curves (f = 10 Hz, R = 0.05 and T = 25C) of samplespresenting the selected surface preparations
S-N curves (R = 0.05, T = 25C)
120
130
140
150
160
170
180
190
200
210
220
230
240
1,00E+05 1,00E+06 1,00E+07
Number of cycles
/
2
(
M
P
a
)
Turned T1
Turned T2
Turned T3
Turned T4
Polished P1
Ground G
Sandblast SB
Polished and sandblast
Turned
Ground
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers33
LMPM
S-N curves (f = 10 Hz, R = -1 and T = 25C) of samples presenting some selected surface preparations
S-N curves (f = 10 Hz, R = -1 and T = 25C) of samples presenting some selected surface preparations
Whler curves (R = -1, T = 25C)
160170180190200210220230240250260270280
1,00E+04 1,00E+05 1,00E+06 1,00E+07Number of cycles
/
2
(
M
P
a
) Turned T1Turned T4Polished P1Ground G1
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers34
LMPM
- A strong influence of the surface finish on the fatigue limit at R = 0.05 and T = 25C:Ground samples ( 125 MPa) < Turned samples ( 155 MPa) < Polished samples ( 195 MPa)
-effects of the surface finish are less pronounced at R = -1 Ground samples ( 175MPa) < Polished samples ( 195 MPa)
A non-influence of the mean stress was commonly assumed for austeniticstainless steels.
Are the results influenced by performing fatigue tests underload control?
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers35
LMPM
INTHERPOL results
100
1000
10000 100000 1000000
Number of cycles
S
a
l
t
(
M
P
a
)
Courbe RCC-Mcourbe moyenneESSAI1 Knu_minEssai4 dlardageEssai4 cote 135 mm (bross)Essai4 cote 150 mm (bross)Essai4 cote 175mm (meul)Essai2 soudureEssai2 PmaxEssai3 soudureEssai3 Pmax
taper
current area
raw
tournsoft grinded
polished
F. CURTIT, EDF R&D / MMC
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers36
LMPM
Gradual elongation observed at max = 410 MPa
020406080
100120140160180200220240260280300320340360380400420
0 0,009 0,018 0,027 0,036 0,045 0,054 0,063 0,072 0,081 0,09elongation
I
n
i
t
i
a
l
a
p
p
l
i
e
d
v
a
l
u
e
F
/
S
i
n
i
(
M
P
a
)
Gradual elongation observedat R = 0.05 (mean = 215 MPa) and f = 10 Hz
F
/
S
0
a
p
p
l
i
e
d
(
M
P
a
)
elongation
Allongement des prouvettes cycles m = 215 MPa (max = 410 MPa) f = 10 Hz
Allongement
F
/
S
0
a
p
p
l
i
q
u
e
(
M
P
a
)
maximum applied stress(MPa)
elongation (%)after 600
cycles
elongation at failure or at the fatigue limit (%)
420 10.7 11.9 (P1, failure at 289 000 cycles)
410 9.4 -
400 8.8 -
380 5.3 -
350 3.8 -
330 2 2.9 (T3, failure at 572 000 cycles)
310 1.4 -
280 - 1.2 (G, failure at 806 000 cycles)
260 - 0.5 (G, un-failed at 10 millions of cycles)
progressive elongation during the first cycles then accommodation-specimens present different levels of pre-deformation according to the
applied load level
Fatigue tests at R = 0.05 (T= 25C)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers37
LMPM
-Elongation is very low at the fatigue limit level for ground specimens.-In this case it can be concluded about the effect of the mean stress (with regard to theresults at R= -1): a detrimental effect of a positive R is revealed.
Influence of the mean stress on the fatigue limit (T= 25C)
Sample preparation
Fatigue limit (MPa) R = -1
Fatigue limit (MPa)
R = 0.05 (elongation)
Polishing 200 185 (5.3%) Sandblasting - 185 (5.3%)
Turning 190 / 200 150 / 155 (2.5%) Grinding 170 125 (0.5%)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers38
LMPM
Diagram of Haigh
Evolution of the fatigue limit versus the applied mean stress
100110120130140150160170180190200210
0 25 50 75 100 125 150 175 200mean stress (MPa)
/
2
/
2
/
2
/
2
(
M
P
a
)
tat politat meulpo
Fatigue limit of ground specimens is decreased by 50 MPa when a high positive mean stress is applied
Fatigue limit of polished specimens isweakly modified by the applicationof a mean stress
In fact :
The detrimental influence of a positive mean stress on polished specimens is
compensated by the hardeninginduced by cycling the materialunder load control
polishedpolishedground
Additional results for other mean values
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers39
LMPM
- great influence of the surface finish at R = 0.05 and T = 25C :Ground samples ( 125 MPa) < Turned samples ( 155 MPa) < Polished samples ( 195 MPa)
-effects of the surface finish are less pronounced at R = -1 :( 175 MPa/195 MPa)
Validity of the reference curve for polished specimens since establishedunder load control and high mean stresses?
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers40
LMPM
Such results led to reconsider HCF data of austenitic stainless steels
In particular the role of control parameter in fatigue tests hasbeen reexamined :
In most conditions, strain rather than stress amplitude has to be imposed in order to be more representative of the real mechanical solicitations in components
L. Vincent et al, CEA (DMN SRMA ) confirmed the detrimental effect of a mean positive stress for 304L polished specimens
by performing tests under constant
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers41
LMPM
160
170
180
190
200
210
220
0 50 100 150 200 250
Nf>5.106 cyclesNf
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers42
LMPM
Related effect of a monotonic or a cyclic predeformation and the application of a positive mean stress
Related effect of a monotonic or a cyclic predeformation and the application of a positive mean stress
detrimental effect of prehardening in strain
controlled tests particularly with a positive mean stress
beneficial effect under stress control
Courbes de Whler des tats pr-crouis (R = 0.05, T = 25C)
150
160
170
180
190
200
210
220
230
1,00E+05 1,00E+06 1,00E+07Nombre de cycles
(
/
2
)
0
a
p
p
l
i
q
u
e
(
M
P
a
)
Etat P1pr-croui
Etat Mpr-croui
Results are mainly affected by the parameter of control used for the fatigue
tests
[V. Doquet, S. Tahiri]
[S. Petitjean, J. Mendez]
Results recently confirmed on monotonically pre strained specimens by EDF/ECP PhDThesis
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers43
LMPM
Relative role of roughness, surface hardening and residual stresses:
Thermal treatments were applied at different temperature levels to relax residual stresses or even to eliminate the hardened surface layer.
Main results:
- a poor influence of residual stresses which is explained by the relatively high plasticity level reached in 304L even for HCF
- a limited effect of surface hardening .
- a predominant effect of the surface topography:
the acuity, length, depth and orientation of the machining grooves are the important specific features with regard to fatigue damage. Therefore, conventional roughness parameters (Ra, Rt ) are not relevant to predict the fatigue behavior of machined or surface treated samples
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers44
LMPM
Damage processes and fatigue limit
For polished specimens: no cracks are observed on non-failed specimens after 107 cycles.
Fatigue limit is in this case clearly associated with non initiation conditions.
In ground specimens the number of cycles to form a crack of the size of the groove is very similar to the number of cycles needed to propagate it to failure.
Fatigue limit associated with the effective K threshold obtained for long cracks.
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers45
LMPM
fissures de fatigue
0
500
1000
1500
2000
2500
0 100000 200000 300000 400000 500000Nombre de cycles
l
o
n
g
u
e
u
r
e
n
s
u
r
f
a
c
e
(
m
)
Etat T5 -smax = 330MPa
Etat M - smax= 290 MPa
Amorages multiples et coalescence
Crack growth of natural short cracks in turned or ground specimens (R = 0.05)
Initiation on two connected machiningdefaults (turned sample)
Multiple cracking and coalescence processes (ground specimen)
multi cracking
Evolution of crak length with the number of cycles
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers46
LMPM
Curves established for max = 290 MPa and max = 330 MPa are superimposed to theda/dN-Keffective curvefor long cracks
R=0,05tTurned T5
cycledunder
330 MPa
R=0,05 -Ground cycled
290 MPa
Crack propagation curves R = 0.1 ( CT samples) and R = 0.05 (cylindrical specimens)
- R = 0.05 (cylindrical specimen turned T3 (max = 330 MPa)- R = 0.05 (cylindrical specimen : ground - max = 290 MPa)
Crack propagation curves of 304L at RT
1,E-08
1,E-07
1,E-06
1,E-05
1,E-04
1,E-031 10 100
K (MPa.rac(m))
d
a
/
d
N
(
m
m
/
c
y
c
l
e
)
K (MPa. m )
R = 0,1 -
CT -nominal
R=0,1-
CT-effective
- R = 0.1 (CT specimen) (C. Sarrazin-Baudoux et J. Petit - 2001)
But K is not the accurateparameter when plasticity is
high
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers47
LMPM
Needs: Control the residual HCF properties of ASS in presence of a fatigue predamage due to the occurrence of solicitations of high amplitude (stops and starts, power increase events)
Evaluate the effect of the surface finish
Objectives and Procedures
Influence of a LCF pre-damage on the 304Lstainless steel fatigue limit
Study of the 304L behavior loadedunder LCF conditions
Cyclic behavior
LCF induced damage
Fatigue limit of pre-damaged samples
Damage accumulation
Influence of surface finishEffect of mean stressRelevant physical or mechanical parameters for the prediction of the fatigue life.typically t/2 = 0,3%
In the HCF range
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers48
LMPM
1
2
4
3
Test conditions : t/2 = 0,3% R=-1f=0,333 Hz T=25C
NF = 25 000 cycles
3750 cycles 7500 cycles
Crack initiation 4000 cycles
(polished specimens with hardened surface layer)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers49
LMPM
N = 7 500 cycles (30%de NR) (three different specimens)
Scattering associated with the crack length atthe initiation
Initiation at twin boundaries :80 m < L0 < 150 m
The hardened layer delays crack initiation but accelerates crack growth
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers50
LMPM
Crack growth described by a Tomkins et Wareign type law:
dL/dNp = C . (p/2) . L
L = L0 . exp (C . (p/2) . Np)
Material constantset Na ~ 4 000 cycles
with Np=N - Na
Identification of C and from experimental data :
L0 = 80 mp/2 = 0,175 %Lf = 4 500 m
C = 6,5 .10-3 = 2
L = L0 . exp (6,5.10-3 . (p/2)2 . Np)
L
n
(
L
/
L
0
)
Polished specimens with hard surface layer(200 m)
Nombre de cycles en propagation
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers51
LMPM
Effect of surface grinding on LCF resistanceEffect of surface grinding on LCF resistance
polishedsevere grindingspherical grinding
Polished NF 25 000 cycles
Severe grinding NF 7 500 cycles
Spherical grinding NF 20 000 cycles
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers52
LMPM
304L surface, ground with a spherical grindstone
The groove orientations are randomly distributed on the surface.The size of grooves, perpendicular to the loading direction in which cracks preferentially initiate, are in this case similar to the microstructure dimensions.
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers53
LMPM
496621000 cycles
785311000 cycles
267951000 cycles
425101000 cycles
depth (m)Surface length (m)Pre cycling
multi-cracking initiation along one straigth groove
N = 1 000 cycles
Weak scattering on themain crack length on four specimens
The main crack already reaches 662 m surface length and 49 m in depth
Effect of a severe grinding on crack initiation and growth
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers54
LMPM
Crack initiation under LCF
Decrease of the fatigue limit
Even if the initiation of a crack is not achieved, precycling can produce a smalldecrease in the fatigue limit of the 304L due to induced cyclic softening
precycling
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers55
LMPM
Initial crack depth m
F
a
t
i
g
u
e
l
i
m
i
t
M
P
a
Kyh, eff = 2,1 0,4 MPa.m-1/2
K=Y..(pia)1/2
aYK effth
Dpi
,
=
Fatigue limit predicted by the K effective threshold approach
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers56
LMPM
Plasticity threshold 180 MPa
/2 > 190 MPa Crack growth controlled by p/2
/2 < 180 MPa Crack growth controlled by Keff
K effective approch40m
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers57
LMPM
Other subjects
LCF behavior and damage mechanisms in LWR environment
Role of temperatureRole of environment
Always in relation with acceptable surface preparation for industrial components
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers58
LMPM
Environmental effects on fatigue behavior of Reactor Materials in LWR environment.
Many studies have shown that fatigue lives of reactors materials are reduced in LWR environment when compared to those obtained in air at high temperature.
For austenitic stainless steels, fatigue lives in LWR environment depend on 3 key parameters:- Strain rate- Dissolved Oxygen content- Temperature
These parameters are taken into account to evaluate fatigue life correction factor Fen for austenitic stainless steels in reactor coolant environment (expression based on ANL model):
*Fen = Nair(RT) / Neau(LWR) =
*[Chopra et al.; 2007; NUREG/CR-6909]
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers59
LMPMEffects of temperature
The LWR environment effect is enhanced at high temperature.LCF test results obtained on a 304L SS in PWR environment and in Air, at two temperatures 150 C and 300C :fatigue life in water, is strongly reduced at 300C and not so much at 150C even for some LCF tests performed at high strain rates
[Solomon, Amzallag et al.; 2004; PVP Seville]
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers60
LMPM
Effects of environment
For Austenitic Stainless Steel, comparison of LCF test results obtained in air at 300C and in deaerated water at high temperature (sa me strain rate and strain amplitude) :
- Same cyclic stress behavior.- Decreasing fatigue life in water environment.
[Cho et al.; 2008; Materials Science and Engineering (248-256)]
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers61
LMPMEffects of strain rate
[Chopra et al.; 2007; NUREG/CR-6909][Cho et al.; 2008; Materials Science and Engineering (248-256)]
Detrimental effect of decreasing strain rate in deaerated water at high temperature or in air at 300C :
- Increases cyclic stress behavior.- Decreases fatigue life.
Austenitic stainless steels exhibit Dynamic Strain Aging in the temperature range of 200 to 800C.
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers62
LMPMEffects of strain rate
For any strain amplitude, in high temperature deaerated water, reduction of fatigue life increases with decreasing strain rate.
So, decreasing fatigue life in LWR environment is enhanced by low strain rate.
Need for a better consideration of the material behavior in air at high temperature
[2]
[Chopra et al.; 2007; NUREG/CR-6909]
In air, steels, and particularly, austenitic stainless steels are sensitive to Dynamic Strain Aging in the temperature range of 200 to 800C.
*Saturation of environmental effect in High temperature water at :
- 0.0004%/s for 304L SS.- 0.004%/s for 316L SS.
*[Chopra et al.; 2007; NUREG/CR-6909]
**304L, Air 300C0.4%/s & 0.01%/s
xx
[Cho et al.; 2008; Materials Science and Engineering (248-256)]**[De Baglion, Mendez et al.; 2008; PhD AREVA NP in progress]
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers63
LMPM
Tests with a more representative signal : SISVariable strain rates, as close as possible to transients applied to Safety Injection nozzles. The strain history corresponds to a cold shock followed by a hot thermal shock. There are 2 types of signals :
Short SIS (840s) & Long SIS (2400s)
In PWR water, for a representative SIS loading signal and ground specimens, experimental Fen are
much lower than expected Fen penalty factors.
[Le Duff, Lefranois, Vernot et al.; AREVA NP; 2008; PVP Chicago]
Combined effects of PWR environment and surface finish for LCF test performed using a complex representative loading signal
[Le Duff, Lefranois, Vernot et al.; AREVA NP; 2008; PVP Chicago]
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers64
LMPM
Current research work
Determining the mechanisms of crack initiation and growthin vacuum, air and PWR environment
taking account for representative loading signals and temperature effects
Current research AREVA NP ENSMAL. de Baglion PhD Thesis
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers65
LMPM
Other new challenges for the next years
Determine crack initiation and crack growth at the micro and meso scales (scale of microstructure) for austenitic 304L or 316L stainless steels taking account for surface finishes
Flat part
At a macroscopic scaleGlobal texture of 316L :
-RX analysis- EBSD analysis
23% 16% 6% 55% (essentially )
Ex: role of the crystallographic texture
200 m
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers66
LMPM
316L -20C - p/2=2.10-3 - 5000 cycles - Air
D4 (1
-11)[-101]
D4
= 53 DG4
= 47
A6 (-111)[110]
A6
= 104
DG6
= 90
D6
(1-11)[1
10]
D6 =
64 DG6
=
42
[110]
[110]
[-101]
Identification of active slip systems and crack initiation conditions
through EBSD analysis
Schmid factor:
{111} = (nPG . ) ( dg . ) {111} = cos . cos
nPG
)
)
surface
Plan de
g
lissemen
t
) s urface
)direction deglissement
ns urf.
10 m
orientation PGOf slip systems with regard to
the free surface :
cos PG = nPG . nsurf.
At a microscopic (one grain) or mesoscopic scale (several grains)
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers67
LMPM
First stages of crack propagationTransgranular crack initiation
transgranular propagation
- =0.461 > 0.41- =52 et =42( )]6045[],5540[
Sitedamorage
joint de macle
=55 [324]
D4 =0,395
A2=0,341
A2/D4=0,86
- =55- Slip activity in both sides
of the grain boundary
B2=0,461=52=42
Sitedamorage
transgranulaire
propagation
Crack initiation in twin intergranular propagation
The local conditions that favor crack initiation are also favorable to their propagation through the surrounding grains
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers68
LMPM
ANR program : AFGRAPFatigue crack initiation in a grain of a polycrystalline aggregate
and propagation in surrounding grains
- Crack initiation criterion : at the scale of the grain taking into account slip activity- Crack propagation criterion : local configuration of grain boundary and of surrounding grains
Modeling of the fatigue crack initiation and of the first stages of propagation in the 316L (influence
of surface properties)Objectives:
Developments based on numerical and experimental tools :- Discrete Dislocation Dynamics simulation- Crystalline Plasticity theory and calculation applied to polycrystalline aggregates- Experimental identification by micro structural analysis at different scales
Industrial and academic partners :EDF R&D, AREVA NP, CEA (SRMA), ARMINES (ENSMP-ParisTech),LMSSMat of ECP (Paris), SIMAP and SYMME of INPG (Grenoble), LMPM of ENSMA (Poitiers)
coupling
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers69
LMPM
Example of numerical simulation of stress localisation induced by
roughness
Virgin material Pre-strained polished material
Pre-strained and rough
surface
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers70
LMPM
THANK YOU FOR YOUR ATTENTIONTHANK YOU FOR YOUR ATTENTIONTHANK YOU FOR YOUR ATTENTIONTHANK YOU FOR YOUR ATTENTION
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers71
LMPM
Questions to solveQuestions to solve Multi-frequential
Variable amplitude Loading
Welds(Geometric singularities,
Residual StressesHAZ)
Industrial surfacefinish
Bi-axialStress State
Mean static loading (pressure and piping constraints)
Crack network(interaction between cracks)
Wall thicknessStress gradient
Multi disciplinary
18-20 Novembre 2008 EDF / MAI - ENSMA Poitiers72
LMPM