FATIGUE BEHAVIOUR OF LOW-ALLOY FERRITIC STEEL AND ...

17th International Conference on Environmental Degradation of Materials in Nuclear Power Systems – Water Reactors August 9-13, 2015, Ottawa, Ontario, Canada

1

FATIGUE BEHAVIOUR OF LOW-ALLOY FERRITIC STEEL AND CORRESPONDING DISSIMILAR METAL WELD SUBJECTED TO OXYGENATED

HIGH TEMPERATURE WATER

M. C. Kammerer, X. Schuler, K.-H. Herter, T. Weissenberg, S. Weihe

Materials Testing Institute (MPA) University of Stuttgart,

Postfach 80 11 40, Stuttgart; 70569, Germany

E-Mail: [email protected]

ABSTRACT

Pressurized components in nuclear power plants (NPPs) operate in high temperature water environments. Deficiencies in considering environmental effects on fatigue in common design standards for nuclear components are controversially discussed. In order to provide a general basis for fatigue design that accounts for environmental effects on fatigue, NUREG CR-6909 provides revised fatigue design methodology where environmental effects from exposure to a reactor coolant environment are considered as a function of strain-amplitude, strain-rate, material type, temperature, and the oxygen content of the water environment. Notwithstanding the experimental basis for the NUREG CR-6909 methodology, there is international consensus that the approach is overly conservative and there is also consensus that there is much scope to optimize the methodology, particularly by the targeted generation of additional data. The current paper reports progress in obtaining better environmental fatigue data for a low alloy ferritic steel (20MnMoNi5-5) used in German NPPs, and for a representative dissimilar metal weld. Following a common methodology, both air and high temperature water fatigue testing is conducted to obtain a specific environmental reduction factor for different loading conditions.

In addition, the formation of, and mechanical behavior of oxide layers have been studied in order to identify and understand the basic mechanisms of environmental fatigue and develop a mechanistic model. Significant conclusions and observations are stated.

Keywords: corrosion-fatigue, environmental effects, low-alloy steel, dissimilar metal weld, oxygenated water, strain-controlled, strain rate

1. INTRODUCTION

Fatigue assessment is an important aspect within the ageing management of safety relevant components in nuclear power plants (NPPs). For pressure retaining components of NPPs thermal loadings are the major cause of cyclic loading. Thermal transients usually result from changes in operational conditions, such as start-up and shut-down procedures related to periodic maintenance. Short term variations in loading resulting from fluctuating power output also lead to increased numbers of fatigue cycles. Locations where stress or strain localisation can occur need special attention [1].

A variety of national fatigue design procedures is available, including a variety of assessment methodologies but all are based on material specific fatigue data depending on the materials being used in NPPs components. For German NPPs, fatigue assessments follow the KTA-standard [2]. The KTA assessment is comparable to ASME-CODE Section III [3]. Neither the KTA, nor the ASME assessments explicitly account for environmental effects on fatigue


2

behaviour. Investigations have shown that for specific combinations of loading and environmental conditions, design curves, as included in common codes and standards, may not be conservative, even for austenitic corrosion resistent materials [4]. In order to incorporate environmentally assisted fatigue (EAF) effects in the fatigue assessment given by ASME-Code Sect. III, an extensive proposal by Chopra was published first in 2007 [5] and revised in draft in 2014 [6]. Another important development was publication of ASME-Code Case N-761 that provided fatigue design curves for light water reactor environments [7].

Comparable investigations to the work by Chopra et al. have been performed at MPA Stuttgart for specific austenitic stainless steels used in German NPPs in two test campaigns [8] and [9]. Environmental fatigue tests with polished round specimens in oxygenated high temperature water were performed and the results showed a marked decrease in fatigue life compared to air environment tests. However, reductions in fatigue life were less than expected based on data for austenitic stainless steels summarized in [6]. The mean data curves for the German steels were significantly less affected compared to data given in [6]. Most significantly, the number of cycles in to end of life in environmental fatigue tests did not fall below the ASME-design curve. This trend was confirmed by recent research at other institutes [10].

In addition to the testing of austenitic stainless steels, tests have been done by MPA Stuttgart on a ferritic material, a reactor pressure vessel (RPV) low alloy steel (LAS) 22NiMoCr3-7 used in German NPPs [11]. The results were only in partial agreement with estimations for fatigue life proposed by ANL [5] and [6]. Discrepancies were found concerning strain rate dependencies where the German material investigated was found to fail at lower cycles for lower strain rates, compared to predictions using the ANL methodology for air environment.

In order to improve the data base for fatigue behaviour of the LASs used in German NPPs a variety of experimental investigations in both air and in a high temperature water environment have been performed. The current paper reports progress in obtaining better environmental fatigue data for low alloy ferritic steel (20MnMoNi5-5) used in German NPPs and for a representative dissimilar metal weld. Following common methodology, both air and high temperature water fatigue testing is conducted to obtain a specific environmental reduction factor for different loading conditions. In addition, the formation of, and mechanical behavior of, oxide layers have been studied in order to identify and understand the basic mechanisms of environmental fatigue and develop a mechanistic model.

2. EXPERIMENTAL INVESTIGATIONS

2.1 Materials

The materials investigated are a ferritic low alloy steel 20MnMoNi5-5 (Mat.-Nr.: 1.6310) comparable to SA 533 Grade B Class1, a stabilized austenitic stainless steel X6CrNiNb18-10 (Mat.-Nr.: 1.4550) comparable to AISI 347 stainless steel and a Ni-base weld filler material NiCr70Nb comparable to alloy 182. The chemical composition is given in Table 1 for the ferrite material, in Table 2 for the austenitic material and in Table 3 for the Ni-base weld filler material. While austenitic and ferrite base materials meet the requirements very well, the chemical analysis of the Ni-base weld filler materials NiCr70Nb shows some deviation to the required composition. This may be due to the fact that specimens used for chemical analysis of the weld material were extracted from a welded section where some material mixing between weld and base materials can have occurred (dilution effects).


3

Figure 1: Geometry of fatigue test specimens

3. FATIGUE TESTS

Fatigue tests were performed using round specimens with a diameter of 10 mm, Figure 1. Tests were performed under strain-control with fully reversed loading sequence (R = -1). Test conditions, in terms of strain amplitude, strain rate and temperature have been chosen such that the effect of the environment on fatigue life may be significant [6]. Fatigue tests were performed in air and high temperature water (HTW) environments with otherwise identical test conditions. Strain-rate can affect the fatigue life in both air and corrosive-media environment, for this reason, air and HTW testing conditions must be identical [11]. The test procedure for HTW includes a load free pre-autoclaving period of at least 100 hours corresponding to the initial operation procedure for new components in NPPs.

260 mm

R10 30° 3

R10

8 0,1

Detail

109,9

10

20


4

Table 1: Chemical composition of ferritic low alloy steel 20MnMoNi5-5

Datasource mass in %

C Si Mn P S Cr Mo Ni Al Cu V Sn N As

MPA

BMWi 1501459A 0.207 0.23 1.27 0.011 0.001 0.05 0.451 0.67 0.025 0.011 0.001 0.002 0.006 0.002

KTA 3201.1

[2]

min 0.15 0.10 1.15 - - - 0.40 0.45 0.010 - - - - -

max 0.25 0.25 1.55 0.012 0.012 0.20 0.55 0.85 0.040 0.12 0.020 0.011 0.013 0.025

Table 2: Chemical composition of austenitic stainless steel X6CrNiNb18-10


C Si Mn P S Co Cr Mo Nb Ni Ti

MPA BMWi 1501459A 0.046 0.39 2.04 0.021 0.004 0.057 17.2 0.288 0.630 10.34 0.007

KTA 3201.1 [2]

min 17.0 10x(C))

max 0.04 1.0 2.0 0.035 0.015 0.2 19.0 - 0.65 12.0 -

Table 3: Chemical composition of Ni-base weld alloy NiCr70Nb


C Si Mn Cr Ni Cu Nb Ti Fe

MPA - BMWi 1501459A 0.03 0.45 5.3 19.4 66.0 0.06 2.4 0.09 5.8

DIN-EN-ISO 14172

min - - 2.0 18.0 63.0 - 1.5 - -

max 0.10 0.08 6.0 22.0 0.5 3.0 0.5 4.0

The chemical composition of the HTW is controlled by a water treatment facility providing consistent and constant boiling water reactor (BWR) water chemistry conditions, of 400 ppb dissolved oxygen and an electrical conductivity < 0.15 µS/cm. Previous tests have shown that stress-corrosion-cracks (SCC) can be caused by placing clip-gage tips on the test-diameter of the specimen surface in corrosive environments. The SCC associated with the contact points leads to significantly lower number of cycles to failure in a fatigue test. Consequently, strain-measurements in high-temperature water are based on a contact-free measuring technique within the autoclave. The autoclave system consists of an air, and a water autoclave which are geometrically very similar. Different heating mechanisms are used due to the fact that in air environment a lower heat transfer coefficient requires more heating capacity. To be able to measure the strain within the testing diameter by use of a clip-gage in air environment a fully sealing of the autoclave is not applicable, Figure 2.


5

Figure 2: Test setup for fatigue testing

First a set of calibration tests in air environment was conducted. Strain-control is realized by use of local clip-gage-measurement. In addition an integral strain measurement is carried out by use of a linear-variable-amplitude-transformer (LVDT), Figure 2. Based on a calibration curve correlating local displacement in the area of the clip-gage, an integral displacement between the LVDT-positions on top and at the bottom can be established. Fatigue tests are performed by use of the integral displacement using the calibrated correlation function. Applying the described testing-methodology, identical test conditions for air and HTW environment are achieved. Finally, an environmental fatigue life reduction factor FEN following the methodology given by Chopra [6] will be derived, Eqn. 1. With respect to other investigations [12] it seems reasonable to derive FEN from fully identical test conditions (temperature, strain rate) in air and high temperature water environment only.

waterNairN

ENF Eqn. 1

In order to extend the knowledge base for different fatigue life influencing factors in air and HTW environment a variation of different test-parameters has been realized in the test setup, Table 4.

3.1 Low alloy steel (20MnMoNi5-5)

The specimens were taken from a pipe segment 800 mm in axial direction, Figure 3. The circumferential distribution of strength properties was found to be homogeneous. The specimens were cut, turned and polished to achieve a surface quality with a mean surface roughness of below 1.5 µm.

medium: airclip-gauge + LVDT

medium: HTWLVDT

LVDT

Clip-gauge

specimen

autoclave

17th InternAugust 9-1

Figure 3:

Table 4:

M

20Mn

NiC

3.2 D

The disaustenithighly vof the wpipes ofmean su

national Confere13, 2015, Ottaw

: Location of

Test conditio

Material

nMoNi5-5

Cr70Nb

issimilar m

similar mettic stainless varying strenwelded sectif LAS 20Mnurface rough

ence on Environwa, Ontario, Can

f fatigue speci

ons for fatigue

T

[°C] [

150 0

240 0

240 0

240 0

240 0

metal weld

tal weld consteel X6Cr

ngth properion, Figure nMoNi5-5 thness below

nmental Degradnada

imens within

e tests

a

%/s]

[%/s]

0.25 0.04

0.25 0.04

0.25 0.4

0.35 0.02

0.30 0.04

nsists of LArNiNb18-10rties the gau5. To be co

the welded sw 1.5 µm.

dation of Materia

800 mm fe

Fatigue tes

Air /HTW

3 / 3

3 / 3

3 / 3

3 / 3

3 / 2

AS 20MnM0, Figure 4. uge section oonsistent wispecimens w

als in Nuclear P

errite pipe (le

Test condition

ts

W

Averag

A

448

602

980

87

221

oNi5-5, Ni-Since the d

of the specimith the specwere also fab

Power Systems

eft)

ge Fatigue Life

Air / HTW

865 / 14954

216 / 9116

017 / 35559

774 / 1917

184 / 16317

-Base weld different typmens was fucimens madebricated wit

– Water Reacto

Environme

FE

3

6

2

4

1

alloy NiCrpes of mateully placed ie from homth a surface

ors

6

ental Factor

EN

.0

.6

.8

.6

.4

r70Nb and erials have in the area

mogeneous quality of


Figure 4:

As a reSpecimebeen inc

Figure 5:

ferrit20MnMo


: Composition

sult of the ens where mcluded in th

: Location of

teoNi5-5

buNi


n of dissimila

welding prmajor defecte fatigue tes

f fatigue speci

ufferingCr19Nb

nmental Degradnada

ar metal weld

rocess local ts like porest.

imens within

weldNiCr19Nb

dation of Materia

defects ares were mac

300 mm w

ba

X6C

als in Nuclear P

e more likeroscopically

welded pipe

austeniteCrNiNb18-10

Power Systems

ly to appeay visible on

– Water Reacto

ar within a n the surface

ors

7

specimen. e have not


8

4. EXPERIMENTAL RESULTS

4.1 Deformation behaviour in Air

The cyclic deformation behaviour is mainly characterized by cyclic hardening or softening during cyclic loading of a material. Generally soft-annealed materials tend to undergo cyclic hardening and cold worked materials tend to show cyclic softening. For piping materials which are primarily loaded by thermal transients small strain rates and comparatively small amplitudes are usually appropriate for fatigue life assessment.

Figure 6: Stress-strain-hysteresis of a stabilized cycle for different temperatures and strain rates in air environment

4.2 Low alloy steel (20MnMoNi5-5)

The deformation behaviour of LAS 20MnMoNi5-5 was found to be sensitive to strain rate and temperature, see Figure 6. For a strain amplitude of a = 0.25 %, a strain rate of ε = 0.04 %/s gave slightly greater hardening compared to a strain rate of ε = 0.4 %/s. Also comparing test temperatures 240 °C to 150 °C, hardening is slightly greater at the higher temperature. This trend is in good agreement with test results of similar LAS tested at MPA Stuttgart [11] but opposite to the behaviour described by ANL [6].

4.3 Dissimilar metal weld

The specimens made from the dissimilar metal weld are tested at a = 0.30 % and show a minor initial hardening behaviour from 400 MPa to 425 MPa. Although the welded section is a highly inhomogeneous built up section as result of the high number of weld passes, the fatigue behaviour is found to be very homogenous with a small scatter in the stabilized stress-strain-relation, Figure 7.

-0,4 -0,3 -0,2 -0,1 0,0 0,1 0,2 0,3 0,4

-400

-200

0

200

400

stre

ss

[MP

a]

strain [%]

a = 0,25%, = 0,4%/s (stabilized), T = 240 °C

a = 0,25%, = 0,04%/s (stabilized), T = 240 °C

a = 0,25%, = 0,4%/s (stabilized), T = 150 °C


Figure 7:

5. FA

Failure stabilizeover moin Table

5.1 L

BesidestemperaincreaseClearly needs tenvironmstrain raan envitemperaenvironmNUREGFEN = 2.amplitudamplitud


: Stress-evolu

ATIGUE LI

criterion toed behaviouore than 80 e 4.

ow alloy ste

hardening ature in air e in fatigue

this discrepto considermental factoate effect inironmental ature from ment, but i

G 6909 [6] .0. Within thdes, a = 0.3de.


ution in three

FE IN AIR

o evaluate fur. All mater% of the te

eel (20MnM

behaviour environmenlife was foupancy demor material or. This stan

n air environfactor FEN 150 °C to

inversely, aa lower bouhis research35 % comp

nmental Degradnada

strain contro

R AND HTW

fatigue life rials investi

esting period

MoNi5-5)

also fatiguent. For an inund to be inonstrates th

specific snds in contrnment is no

of 10.8 in240 °C res

a decreaseundary for

h FEN was fopared to a =

dation of Materia

olled fatigue t

W

was a loadigated exhibd. Results fo

e life was foncrease in l

n the range ohat in order strain rate rast to what ot considerenstead of osulted in ain cycles toenvironmen

ound to be 3= 0.25 %, F

als in Nuclear P

tests of welde

d drop of 2bited distincfor LAS and

found to be loading rateof 1.6 for air

to derive vsensitivity

is publishedd for LAS.only 6.6, ian increaseo failure inntal effects .0 at 150 °C

FEN tends to

Power Systems

d specimens

25 % basedctive stabilizd welded sp

sensitive toe by a factor and 3.9 fovalid enviroy to yield d in NUREGNeglecting

s calculatedin cycles

n HTW envat T 150

C. Regardio decrease w

– Water Reacto

d on the stezed fatigue

pecimens are

o both strainor of 10 at 2or HTW envonmental fa

a physicaG 6909 [6], g the strain rd. A changto failure

vironment. °C is specng highe

with increas

ors

9

eady state behaviour e included

n rate and 240 °C an vironment. actors, one ally valid where the rate effect ge in test in an air Based on

cified with er strain sing strain


10

5.2 Dissimilar metal welds

Before reporting results of the fatigue tests on weld metal, it is important to note that no specimens with known weld defects were included in the program. Also, none of the initially surface-defect-free specimens failed prematurely due to buried weld defects influencing. Fatigue life for dissimilar metal welds in air environment at 240 °C at a strain amplitude of a = 0.30 % with a strain rate of 0.04%/s was determined to be in average at about 22184 loading cycles, Table 4. For HTW environment fatigue life was found to be at about 16121 loading cycles. Based on NUREG 6909 [6] for NiCr-ferrous alloys under the given environmental and loading conditions, FEN is predicted to be 1.3. Within this research FEN was found to be 1.4 for the given conditions.

6. OXIDE LAYERS IN HTW

Oxide layers have been studied for various types of materials in combination with different types of environment like air, water or steam in numerous research projects. In steam environments, oxide layers can grow to thicknesses up to several tenths of millimeters, depending on the materials, the chemical composition of the environment and the mechanical loading scenario, [13] and [14]. Such layers are comparatively thick and can be investigated in cross-section by means of optical microscopy and scanning electron microscopy [13]. In contrast, the thickness of oxide layers grown on ferritic steels in oxygenated HTW are thinner and characterization requires electron microscopy, and the capabilities of electron microscopy are greatly enhance if combined with a focused ion beam attachment to facilitate specimen extraction and preparation. Oxide layers on low alloy or carbon steels are generally known to consist of an inner fine grained and outer coarse grained oxide layer, Figure 8. The inner oxide layer grows along the former metal grain structure and is comparatively dense. The outer oxide layer grows on top of the former metal surface and is fed by diffusion processes of hydrogen ions being transported into the base metal and metal-ions being dissolved out of the base metal structure. In oxygenated HTW environment both oxide layers mainly consist of oxide-species magnetite which is formed by Schikorr-reaction [15]. The properties of protective oxides on ferritic steels are thought to be influential in their cracking behaviour in HTW environments. In general mechanical behaviour of oxide layers is very brittle compared to the metallic base material. Experimental investigations presume that threshold strains for cracking of protective oxide layers exist [16],[17] and [20]. The values taken from literature exhibit a big scatter for critical strains and it may be concluded that the threshold values are very sensitive to both test conditions and the material.


Figure 8:

6.1 Fo

Investigpronounoverviewin air an

The chastructurecharacteattachedtests sotopologyAt the escanning

The surshown isomewhstrain raform les

In additthe specstrain rawith ε =visible strain raduring fdeforma


: Composition

ormation

gations focunced and ow and compnd HTW env

aracteristics es, by viewerizing oxidd to an electome plain sy and end of the fg electron m

rface topoloin Figure 11hat finer andate of ε = 0.ss pronounc

tion to topolcimens withate of ε = 0.0= 0.4 %/s. Bin the case ate. Obvioufaster cyclination behav


n of oxide lay

ussed on theoxide layersparison of thvironment is

of oxide laywing in crode thicknestron microsspecimens w

the FIfatigue testsmicroscopy.

gy of the fe1 for differed magnetite4 %/s the o

ced edges.

logical analh strain amp04 %/s the o

Based on theof the low

usly the grang is signifiviour of th

nmental Degradnada

yer on low allo

e ferritic mas are expeche different s given in F

yers were inoss-sectionss and its lcope was uwere only pB-Cut ths (25% load

errite speciment strain rate grains seeoverall magn

lysis, FIB-cuplitude a = 0overall oxidese results a

wer strain raain size of icantly bigg

he ferritic b

dation of Materia

oyed steels in

aterial whercted to be types of ox

Figure 9.

nvestigated using a sc

layered struused to prepapre-autoclavhrough ld drop) the

mens in HTtes. At ε = 0m less mergnetite size is

ut cross-sec0.25 % at dde layer thica distinctionate but can the magnet

ger. These rebase metal

als in Nuclear P

n HTW enviro

re environmmechanica

xide layers o

with regardcanning eleucture, a foare cross seved and aflayers arspecimens

TW tested a0.04 %/s theged compars topologica

ctions were different strackness is clen between in

approximattite grains gesults indica

may direc

Power Systems

onments

mental effectally less staon the differ

d to the surfectron microocused ion ections. In afterwards anre shownwere sectio

t strain ampe magnetite red to ε = 0ally bigger a

made and eain rates, Figearly thinnernner and outely be derigrown on tate that the ctly influen

– Water Reacto

ts are generable. A maent types of

face topologoscope. Tobeam (FIB

addition to tnalysed. Thn in Foned and an

plitude a = grain struc

0.4 %/s. At tand magnet

examined fogure 12. At r compared

uter oxide laived from tthe outer oxstrain rate d

nce formatio

ors

11

rally more acroscopic f materials

gy of grain facilitate

B) facility the fatigue he surface Figure 10.

nalysed by

0.25 % is ture looks the higher ite spinels

or some of the lower to the test

ayer is not the higher xide layer dependent on of the


magnetithan whlayers oaveragebase wepredomimixing cracks e

Figure 9:

6.2 M

Values large rantype of failure” and [20]0.08 %

Within tension acousticdetectedacousticmagnetiaustenittension did not thicknesmicroph


ite oxide layhat is knownon the austen thickness.

eld material,inantly buil(dilution ef

exist, Figure

: Dissimilar m

Mechanical b

for stress ornge and seecomponentdata for f

]. In terms [20].

this researctest on pre

c emission td at about 3c cracking site grains frote and Ni-bclose to thedetect crack

ss comparedhone transm


yer. Overaln from liternite materiaThe oxide l, the oxide llt up of sinffects) durine 13.

Air

metal weld aft

behaviour

r strain limiem to be higt, and its opferritic mateof critical s

ch, mechanietreated spetechnique (A350 MPa resignals correom the innebase materie yield strenking of the d to layers o

mitter.

nmental Degradnada

l, the thicknrature, [18]al were founlayers look layers are mngle small cng the weld

ter fatigue tes

its for differghly dependperational lerials is in strains to fa

ical behaviocimens (HTAET). For telating to 0espond to crer layer coulal was pergth of the auaustenite an

on ferrite m

dation of Materia

ness of the and [19]. I

nd to be verytypical for

mostly similacrystals or ding process

st in air and H

rent types oent on, enviloading situ

the range ailure these

our of oxideTW, 100 hothe LAS ma.15 to 0.17 %racking of tld not be detformed. Austenite mand Ni-basis

material and

als in Nuclear P

oxide layern contrast ty homogenothis type ofar to the ausclusters ofs local mag

HTW

f oxide layeironmental c

uation. Mosof 100 MPwere found

e layers waours, 400 ppaterial, crac% of uniaxthe inner oxtermined. Awelded sp

aterial at abos layers. Thitherefore hi

Power Systems

rs tends to to the ferritious and cleaf material. Istenite mateoxide-cryst

gnetite isles

HTW

ers given in conditions, t of the ox

Pa to 400 Md to be in th

as investigatpb oxygen) cking withinxial strain. Wxide layer orAn analogou

ecimen waout 210 MPais may be digher requir

– Water Reacto

be somewhic steel oxidarly below In the case erial. The sutals. Due to and defect

W

the literatuthe base maide “critica

MPa in tenhe range of

ted by use oin combina

n the oxide Whether ther separation

us investigatas loaded ina. Analysis due to the smrements reg

ors

12

hat thinner des, oxide 100 nm in of the Ni-

urface was o material ts like hot

ure cover a aterial, the al-stress to sion, [14] f 0.06 and

of a static ation with layer was

e detected n of bigger tion on the n uniaxial with AET mall layer arding the


F

Figure 10

I

Figure 11

I

Figure 12


FIB-Cut thr

0: Oxide laye

I. a = 0.25

1: Surface top

I. a = 0.25 %

2: FIB-Cut th


rough oxide

rs of pretreat

%, ε = 0.04

pology of the

%, ε = 0.04

hrough oxide

nmental Degradnada

layers

ted LAS spec

%/s

oxide layers

%/s

layers on the

dation of Materia

cimens withou

formed on th

e LAS after fa

als in Nuclear P

Su

ut cyclic loadi

II. a

he LAS during

II. a

atigue test in

Power Systems

urface topolo

ing

a = 0.25 %, ε

g a fatigue tes

a = 0.25 %,

HTW

– Water Reacto

ogy

ε = 0.4 %/s

st in HTW.

ε = 0.4 %/s

ors

13

s


Figure 13increase


3: Hot crack iron concentr


on the Ni-barations (right

nmental Degradnada

ase weld sectit

dation of Materia

ion (left) and

als in Nuclear P

locally grow

Power Systems

wn big magne

– Water Reacto

tite grains as

ors

14

s a result of


15

7. DISCUSSION AND CONCLUSIONS

Environmental effects on fatigue behaviour of materials for pressure retaining components are investigated to improve the understanding for mechanical-chemical interactions. This paper is focused on a specific ferritic material typical of the type used for German NPPs, together with a representative dissimilar metal weld.

Results demonstrate that, especially in the case of the LAS tested, generic estimations of EAF in terms of fatigue life reduction factors do not represent test results in a satisfactory manner. Based on a common methodology, fatigue life, even in air, can differ by a factor of about 1.5 for the low cycle fatigue regime between 103 and 105 loading cycles, compared to results in literature [6]. In order to evaluate relative-to-air material behaviour in HTW correctly, it is necessary to consider effects of temperature and strain rate in air as well as in a HTW environment to generate accurate, material specific, fatigue life reduction factors. Merging fatigue data for different types of low-alloy steels together into a single generic database, disregards some, possibly significant, material specific variations in EAF behaviour arising from different mechanisms.

In contrast to the results for the ferritic material, our investigations of Ni-base dissimilar metal weld material EAF behaviour produced results in good agreement with values given in the literature [6], showing only a less than 10 % deviation from the expected FEN-factor for NiCr-ferrous alloys [6]. Although chemical analysis clearly detected local concentration increases of iron within the boundary layer of the weld, it seems this does not influence the cracking behaviour significantly.

An analysis of oxide layers on the ferrite material shows a distinct correlation between strain rate and the development of oxide layers. In order to establish a mechanistic model of the interaction between base material and oxide layers it is necessary to extend the experimental database of environmental fatigue tests for different loading conditions. To understand the basic mechanisms, such investigations have to be carried out on very homogeneous materials in order to reveal the subtle interactions, and relatively small scale surface effects that are marginal from a general mechanical engineering standpoint, but may be influential in determining EAF behaviour.

8. ACKNOWLEDGMENTS

The presented results were funded by the German Federal Ministry of Economic Affairs and Energy (BMWi project no. 1501459A) on the basis of a decision by the German Bundestag.


16

9. NOMENCLATURE

EAF Environmentally assisted Fatigue

HTW High temperature water NPPs Nuclear power plants MPA Materials testing Institute

University of Stuttgart BMWi German Federal Ministry of

Economic Affairs and Energy BWR Boiler Water Reactor LWR Light Water Reactor ANL Argonne National Laboratory LVDT Linear Variable amplitude

Transformer LAS low alloy steel AET acoustic emission technique SCC stress corrosion crack FIB focused ion beam strain stress electrical conductivity


17

10. REFERENCES

[1] Krätschmer, D., Roos, E., Schuler, X., Herter, K.-H., Proof of fatigue strength of nuclear components part II: Numerical fatigue analysis for transient stratification loading considering environmental effects, Pressure Vessels and Piping 92 (2010) 1-10

[2] German Nuclear Safety Standard, KTA 3201: Components of the Reactor Coolant Pressure Boundary of Light Water Reactors. Part 2: Design and Analysis, Ed. 11/2013

[3] Rules for Construction of Nuclear Power Plant Components. ASME Boiler and Pressure Vessel Code, Section III, Subsection NB, The American Society of Mechanical Engineers (2013)

[4] Chopra, O. K., Mechanism and estimation of fatigue crack initiation in austenitic stainless steels in LWR environments, NUREG/CR-6787, ANL-01/25, 2002

[5] O. K. Chopra, W. J. Shack, Effect of LWR Coolant Environments on the Fatigue Life of Reactor Materials. Report: ANL-06/08 NUREG/CR-6909, 2007

[6] O. K. Chopra, W. J. Shack, Effect of LWR Coolant Environments on the Fatigue Life of Reactor Materials. Report: ANL-12/60, NUREG/CR-6909), Argonne (Illinois, USA), 2014

[7] Code Case N-761, Fatigue Design Curves for Light Water Reactor (LWR) Environment - Cases of ASME Boiler and Pressure Vessel Code, Approval Date September 2010

[8] Weissenberg, T., Investigation of the influence of the reactor coolant on the fatigue behaviour of austenitic stainless steels, Work package , BMU-Project SR 0801312, MPA University of Stuttgart, 2011

[9] H. Waidele, T. Weissenberg, Work Package 3.1: Influence of the reactor coolant on the fatigue behaviour of austenitic stainless steels. Report: MPA University of Stuttgart, BMU-SR 2501, Stuttgart (Germany), 2007

[10] J. Solin, S. Reese, H. Karabaki, W. Mayinger, Environmental fatigue factors (NUREG/CR-6909) and strain controlled data for stabilized austenitic strainless steel. ASME 2013 PVP, 14-18.7.2013, ASME International, Paris (France), Paper PVP2013-97500

[11] T. Weissenberg, Fatigue behavior of Ferritic Steels for Pressure Vessels and Piping in Oxygenated High Temperature Water. Report: MPA University of Stuttgart, BMWi 1501309, 2009

[12] E. Roos, K. Maile, K.-H. Herter, X. Schuler, Experimental Database to evaluate different parameters influencing the fatigue S/N-Curve, Materials Testing Institute (MPA) University of Stuttgart, The international conference on fatigue of reactor components, October 3-6 2004

[13] H. Schuster, Magnetitbildung und Druckverlustanstieg im Verdampfer von Bensonkesseln, Allianz-Berichte für Betriebstechnik und Schadensverhütung, April 1971

[14] H.-G. Heitmann, Chemie und Korrosion in Kraftwerken, Vulkan-Verlag, Essen 2000

[15] G. Schikorr, Über Eisen(II)-Hydroxid und ein ferromagnetisches Eisen(III)-Hydroxid, Z. anorg. U. allg. Chem. 212, 1933

[16] E. M. Field, R. C: Stanley, A. M. Adams, D. R. Holmes, The Growth Structure and Breakdown of Magnetite Films on Mild Steel, 2nd International Congress on Metallic Corrosion, New York, 1963

[17] P. L. Harrison, Tensile fracture of Magnetite films, Corrosion Science Vol 7(1967), pp. 789 – 795 [18] H. S. Gadiyar, N.S.D. Elayathu, Corrosion and Magnetite Growth on Carbon Steels in Water at

310 °C – Effect of Dissolved Oxygen pH and EDTA Addition, Corrosion 36 p. 306-312, June 1980

[19] K. Mabuchi, Y. Horii, H. Takahashi, M. Nagayama, Effect of Temperature Dissolved Oxygen on the Corrosion Behaviour of Carbon Steel in High Temperature Water, Corrosion 47 p. 500-508, July 1991

[20] M. M. Navab-Motlagh, Das mechanische Verhalten der Magnetitschutzschicht und dessen Berücksichtigung in der Festigkeitsberechnung, PhD-thesis, MPA Universität Stuttgart, 1984

FATIGUE BEHAVIOUR OF LOW-ALLOY FERRITIC STEEL AND ...

Documents

Transcript of FATIGUE BEHAVIOUR OF LOW-ALLOY FERRITIC STEEL AND ...