fatigue behaviour and crack growth of ferritic steel under ......fatigue tests were performed in the...

Fatigue behaviour and crack growth of ferritic steel under environmental conditions

K.-H. Herter, X. Schuler, T. Weißenberg MPA University of Stuttgart, Stuttgart

38th MPA-Seminar October 1 and 2, 2012 in Stuttgart

155

Abstract

The assessment of fatigue and cyclic crack growth behaviour of safety relevant components is of importance for the ageing management with regard to safety and reliability. For cyclic stress evaluation different codes and standards provide fatigue analysis procedures to be performed considering the various mechanical and thermal loading histories and geometric complexities of the components. For the fatigue design curves used as a limiting criteria the influence of different factors like e.g. environment, surface finish and temperature must be taken into consideration in an appropriate way.

Fatigue tests were performed in the low cycle fatigue (LCF) und high cycle fatigue (HCF) regime with low alloy steels as well as with Nb- and Ti-stabilized German austenitic stainless steels in air and high temperature (HT) boiling water reactor environment to extend the state of knowledge of environmentally assisted fatigue (EAF) as it can occur in boiling water reactor (BWR) plants.

Using the reactor pressure vessel (RPV) steel 22NiMoCr3-7 experimental data were developed to verify the influence of BWR coolant environment (high purity water as well as sulphate containing water with 90 ppb SO4 at a test temperature of 240 °C and an oxygen content of 400 ppb) on the fatigue life and to extend the basis for a reliable estimation of the remaining service life of reactor components. Corresponding experiments in air were performed to establish reference data to determine the environmental correction factor Fen accounting for the environment.

The experimental results are compared with international available mean data curves, the new design curves and on the basis of the environmental factor Fen.

Furthermore the behaviour of steel 22NiMoCr3-7 in oxygenated high temperature water under transient loading conditions was investigated with respect to crack initiation and cyclic crack growth. In this process the stress state of the specimen and the chemical composition of the high temperature water play an important role for the transferability to real components. Environmentally assisted cracking (EAC) experiments were performed with fracture mechanics C(T)-specimens of different size in high temperature water autoclaves under simulated BWR water conditions. In some of the experiments the influence of chloride transients on crack initiation and crack propagation was investigated.

In this paper the position of MPA University of Stuttgart concerning material fatigue data, mean data curves and fatigue design curves including environmental effects is demonstrated.

1 Introduction

The basis for construction, design and operation of nuclear systems, structures and components (SSC) are national technical codes and standards like the ASME-Code Section III [1], the French RCC-M Code [2] or the German Nuclear Safety Standards KTA [3]. The basic philosophy in the design of SSC is to demonstrate that the function and the integrity are guaranteed throughout the lifetime. It is important that the design concept accounts for most possible failure modes and provides rational margins of safety against each type of failure mode. Some of the potential failure modes which SSC designers should take into account are for example: Excessive elastic deformation including elastic instability, excessive plastic deformation, brittle fracture, fatigue and corrosion.

During design stage a complete picture of the state of stresses within the SSC obtained by calculation or measurement of both mechanical and thermal stresses during steady state operation and transient loading has to be created. It has to be demonstrated that all stresses (primary, secondary) as well as environmental loading are within the allowable stress limits given by the codes and standards, and the usage factor developed by a fatigue analysis (peak stresses) is well below the limiting value (cumulative fatigue life usage factor U).

156

It is possible to prevent failure modes caused by fatigue by imposing distinct limits on the peak stresses at the highest loaded regions of the SSC or by reducing the load cycles since fatigue failure is related to and initiated by high local stresses respectively strains. The design rules according to the technical codes and standards [1], [2] and [3] provides for explicit consideration of cyclic operation using design fatigue curves of allowable alternating loads (allowable stress or strain amplitudes) vs. number of loading cycles (S/N-curves) specific rules for assessing the cumulative fatigue damage caused by different specified or monitored load cycles. The influence of different factors like welds, environment, surface finish, temperature, mean stress and size must be taken into consideration in an appropriate way.

Fatigue analysis can be performed by different concepts which must not necessarily yield in the same result, Fig. 1. In nuclear codes and standards usually local concepts are used (local notch strain or stress concepts).

Fatigue Strength AssessmentFatigue Strength Assessment

Experimental

Assessment

Experimental

Assessment Analytical / Numerical AssessmentAnalytical / Numerical Assessment

Global Concept

Based on external forces or based on nominal stresses within the component

sections in question

Global Concept



Local Concept

Based on local stresses and related to criticalvalues of local phenomena like crack initiation

or crack propagation

Local Concept



Nominal StressConcept


Structural StressConcept


Local Notch stressConcept


Crack PropagationConcept


Full Scale

Component Test

Tests up to failure of thecomponent under typical

loading conditionsNF=f(loading),

eg. acc. to ASME-VIII, Div.2 or up to the tech-

nical crack initiation Ni=f(loading)

Full Scale

Component Test





� Nominal Stress-Woehler-Line

- Material- Geometry- Surface finish /

Weld quality� Nominal Stress

Collektive- Loading sequence- Geometry- Load amplitude

� Accumulation of damage

NF=f(σNom)

� Structural Stress resp.Strain Woehler-Line


Weld quality� Structural Stress resp.

Strain Collektive- Loading sequence - Geometry- Load amplitude

� Akkumulation of damage

NF=f(σStruc)

� Strain Woehler-line /cyclic Stress-StrainCurve

- Material- Weld quality

� Local Notch Stress- Geometry- Load amplitude

� Stress-Strain History- Load-time function

� Accumulation ofdamage

NI=f(σNotch)

� Crack Growth Curve

(cyclic)- Material- Load amplitude

� Load-Time Function ∆Keff

- Geometry- Crack geometry

� Crack Growth / Lifetime

� Woehler-Line / Service-Life Curves (crack init. Up tofailure, ∆σNom)

Fatigue Strength AssessmentFatigue Strength Assessment

Experimental

Assessment

Experimental

Assessment Analytical / Numerical AssessmentAnalytical / Numerical Assessment

Global Concept



Global Concept



Local Concept



Local Concept











Full Scale

Component Test





Full Scale

Component Test





� Nominal Stress-Woehler-Line


Weld quality� Nominal Stress

Collektive- Loading sequence- Geometry- Load amplitude

� Accumulation of damage

NF=f(σNom)

� Structural Stress resp.Strain Woehler-Line


Weld quality� Structural Stress resp.

Strain Collektive- Loading sequence - Geometry- Load amplitude

� Akkumulation of damage

NF=f(σStruc)

� Strain Woehler-line /cyclic Stress-StrainCurve

- Material- Weld quality

� Local Notch Stress- Geometry- Load amplitude

� Stress-Strain History- Load-time function

� Accumulation ofdamage

NI=f(σNotch)

� Crack Growth Curve

(cyclic)- Material- Load amplitude

� Load-Time Function ∆Keff

- Geometry- Crack geometry

� Crack Growth / Lifetime

� Woehler-Line / Service-Life Curves (crack init. Up tofailure, ∆σNom)

Fig. 1 : Different Methods to perform fatigue analysis

2 Fatigue analysis in nuclear codes and standards

2.1 Fatigue design curves

Reviewing fatigue analyses to be performed for nuclear pressure vessels and piping it becomes apparent that the majority is similar to or identical with those included in the ASME-Code Section III [1], like the German KTA Standards [3].

Fatigue data are generally obtained from unwelded smooth cylindrical specimens which were tested under strain control at room temperature (RT) and in air environment with a fully reversed loading, i.e. strain ratio of Rεεεε=–1, and are plotted in the form of nominal stress amplitude Sa vs. the number of cycles N to failure [4], [5], [6] and [7]. The total strain range ∆εat obtained from the tests is converted to nominal stress range 2Sa by multiplying the strain range by the modulus of elasticity E at test temperature:

157

2

2at

a ESε∆

⋅= (1)

The design curves (S/N-curves) were derived by introducing factors of 2 on stress (Sσσσσ) and 20 on cycles (SN) on the best-fit curves, whichever gave the lowest curve and is meant to account for real effects (“scatter of data and material variability”, “size effects”, “surface finish and environment”, e.g. [7], [8], [9]) occurring during plant operation. All of the pressure vessel and piping fatigue design rules are based essentially on the same approach based on data from primarily low-cycle fatigue (LCF) tests. Conservative S/N-curves are developed and used for the fatigue analysis in conjunction with stress concentration factors Kt or fatigue strength reduction factors Kf to take into account the structural discontinuities in the SSC including welds [8].

The fatigue life is defined as the number of cycles, N25, necessary for the tensile stress to drop 25 % from its peak or steady–state value during test. For a specimen size usually used in fatigue testing, e.g. 5-12 mm diameter cylindrical specimens, this corresponds to a ≈3 mm deep crack [9].

The existing fatigue ε–N data to develop the S/N-curves are categorized by the types of material like austenitic stainless steels, Fig. 2, and carbon steels and low alloy steels, Fig. 3. Therefore most of the S/N-curves given in the codes and standards are to be applied for specific steels (e.g. distinguish between steels of different ultimate tensile strength Rm).

101

102

103

104

105

106

107

108

109

1010

1011

10

100

1000

10000

Auslegungskurve basierend

auf Lastwechselversuchen


auf Lastwechselversuchen

und maximale Mittelspannung


auf Dehnungswechselversuchen

KTA 3201.2 (Regeländerungsentwurf 12/2010)

Ermüdungskurven für

austenitische Stähle

E = 179 000 MPa

SN = 20

SS = 2

Sp

an

nu

ng

sam

plitu

de σσ σσ

a [

MP

a]

Lastspielzahl N

Fig. 2 : Fatigue design curve austenitic stainless steels [3]

Based on experimental ε–N data of the last twenty years fatigue life models for estimating the fatigue lives of these steels in air have been redeveloped at ANL [9] (NUREG/CR-6090) as best-fit or mean data curves of a Langer type equation [4], [5] and [6]. Based on these data new fatigue design curves were developed different from the curves included in ASME-Code Section III (ed. 2008 and earlier) and KTA Standard [3]. The new carbon steel best-fit curve, Fig. 4, is represented by

).(ln.T..)Nln( a 113097510012406146 −ε−−= (2)

and for low alloy steels, Fig. 5, by

).(ln.T..)Nln( a 151080810012404806 −ε−−= (3)

158

where εa is the applied strain amplitude (%), and T is the test temperature (°C). For austenitic stainless steels in air at temperatures up to 400°C the fatigue data are best represented by the best-fit curve, Fig. 6,

).(ln..)Nln( a 112092018916 −ε−= (4)

101

102

103

104

105

106

10

100

1000

10000

Ermüdgungskurve für

Rm = 790 bis 900 MPa

Ermüdungskurve für

Rm < 550 MPa

KTA 3201.2 (Regeländerungsentwurf 12/2010)

Ermüdungskurven für ferritische Stähle

E = 207 000 MPa

SN = 20

SS = 2

Sp

an

nu

ng

sam

plitu

de σσ σσ

a [

MP

a]

Lastspielzahl N

Fig. 3 : Fatigue design curves carbon steels and low alloy steels [3]

Fig. 4 : Fatigue design curve for carbon steels in air, developed from the ANL model [9] and based on factors of 12 on life and 2 on stress

2.2 Effects influencing the fatigue life

The use of the fatigue design curves is restricted in the nuclear codes and standards to a specific maximum temperature below the creep range. Using design fatigue curves it is necessary to adjust the allowable stresses if the modulus of elasticity E at operating temperature is different from that one used for the design curves. In air, the fatigue life of ferritic steels decreases with increasing temperature; however the effect is relatively small, whereas for austenitic steels according to [9] the temperature has no significant effect on the

159

fatigue life. In the design S/N-curves any temperature effects are accounted for in the subfactor for “data scatter and material variability”.

Fig. 5 : Fatigue design curve for low alloy steels in air, developed from the ANL model [9] and based on factors of 12 on life and 2 on stress

Fig. 6 : Fatigue design curve for austenitic stainless steels in air, developed from the ANL model [9] and based on factors of 12 on life and 2 on stress

In the codes and standards there are different and specific requirements concerning the surface finish of components especially for welded regions, for different vessel and piping products and different joints. A special regard to the influence of surface finish depending upon peak-to-valley height Rz is obvious. Depending on Rz for carbon steel, low–alloy steel and for austenitic steels the surface finish would decrease fatigue life. This effect is accounted for in the subfactor for “surface finish and environment”. Stress indices are available for use of the code equations determining the stress amplitudes. This includes also different weld configurations. Fatigue tests with weld metal showed a tendency for shorter fatigue life compared to the base metal [14].

During the last two decades great endeavors have been made to investigate the influence of the coolant environment on fatigue life, e.g. [10], [11], [12], [13], [14], [15]. Today it is

160

generally accepted that the Light Water Reactor (LWR) environment can have a significant impact on the fatigue life of carbon and low alloy steels as well as on austenitic stainless steels and has to be involved in cumulative fatigue life considerations.

The U.S. Nuclear Regulatory Commission (NRC) issued the Regulatory Guide 1.207 “Guidelines for Evaluating Fatigue Analyses Incorporating the Life Reduction of Metal Components Due to the Effects of the Light-Water Reactor Environment for New Reactors” [16]. This Guide is based on the research work and publications by ANL [9]. The ANL fatigue life model for LWR environments includes parameters for the effects of temperature, strain rate, dissolved oxygen content in water and, in case of ferritic steels, sulphur content of the steel. The environmental effects are expressed in terms of an environmental correction factor Fen as the ratio of fatigue life in a RT at air environment to fatigue life in LWR coolant at operating temperature

water

airen N

NF = (5)

also depending on the above-mentioned parameters, which are used to estimate the environmental fatigue usage

∑ ×= ienien FUU (6)

3 Experimental investigations

3.1 Materials and test parameters

For the tests performed at MPA different nuclear grade materials manufactured according to the requirements of KTA safety standard are available:

• Low alloy ferritic steel 22NiMoCr3-7

• Low alloy ferritic steel 20MnMoNi5-5

• Ni stabilized austenitic stainless steel X10CrNiNb18-9 and X6CrNiNb18-10

• Ti stabilized austenitic stainless steel X10CrNiTi18-9

The material strength parameter yield strength and ultimate tensile strength at room temperature are respectively (all mean values)

• 459 MPa and 613 MPa for material 22NiMoCr3-7 (rolled plate)

• 503 MPa and 653 MPa for material 20MnMoNi5-5 (rolled plate)

• 245 MPa and 573 MPa for material X6CrNiNb18-10 (rolled bar)

• 239 MPa and 548 MPa for material X6CrNiNb18-10 (seamless pipe)

• 292 MPa and 564 MPa for material X10CrNiNb18-9 (seamless pipe)

• 240 MPa and 600 MPa for material X10CrNiTi18-9 (seamless pipe)

All test materials fulfilled the requirements according to KTA safety standard, e.g. chemical composition, mechanical properties (yield strength, ultimate tensile strength, elongation at fracture, percentage reduction of area) and absorbed impact energy.

The smooth cylindrical fatigue test specimens are machined from the plate and pipe materials respectively in rolling and in longitudinal directions. Specimen surface was mechanically polished with a surface roughness of Rz<2 µµµµm. The specimens were tested under strain control at RT and elevated temperatures up to 350 oC with fully reversed loading conditions at different strain amplitudes, i.e. strain ratio of Rεεεε=–1.

161

3.2 LCF tests in air environment

3.2.1 Austenitic material

In addition to more than 60 fatigue data available at MPA Stuttgart in total a data pool of 158 fatigue tests at RT for the stabilized austenitic stainless steels X10CrNiNb18-9, X6CrNiNb18-10 and X10CrNiTi18-9 including the data of different laboratories (MPA Darmstadt, E.ON, MPA Stuttgart) is shown in Fig. 7. The data are best represented by the equation (best-fit or mean data curve, procedure according to [17])

).(ln..)Nln( a 136017227066 −ε−= (7)

This equation fits also load controlled tests at 107 load cycles as well as the approximation of the endurance limit based on the ultimate tensile strength.

0,01

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06 1,0E+08 1,0E+10

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Lod ycles / N

Load Controlled (RT)

Data Pool RT

Based on Ult. Tensile Strengt (RT)

Mean Data Curve RT

Fig. 7 : Fatigue life of stabilized austenitic stainless steels at RT in air

0,01

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06 1,0E+08 1,0E+10

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Load Cycles / N

Load Controlled (288°C)

Data Pool Elev. Temp.

Based on Ult. Tensile Strength

Mean Data Curve Elev. Temp.

Fig. 8 : Fatigue life of stabilized austenitic stainless steels at elevated temperatures in air

162

In addition to more than 80 fatigue data available at MPA Stuttgart in total a data pool of 138 fatigue tests at temperatures between 240 oC and 350 oC for the stabilized austenitic stainless steels X10CrNiNb18-9, X6CrNiNb18-10 X10CrNiTi18-9 from different laboratories (MPA Darmstadt, MPA Stuttgart, AREVA) is shown in Fig. 8. The data are best represented by the equation (best-fit or mean data curve, procedure according to [17])

).(ln..)Nln( a 078025528506 −ε−= (8)

This equation covers also load controlled tests at 107 load cycles as well as the approximation of the endurance limit based on the ultimate tensile strength.

3.2.2 Austenitic cladding

To evaluate the fatigue behaviour of the austenitic cladding of pressurized components, the fatigue design curves are usually based on experiments on specimens taken from plates, pipes or bars. Due to the rapid heat transport into the ferritic base material, the cladding has a distinctive anisotropic structure that results from the manufacturing process. Therefore, it may not be assumed a priori that the fatigue life curves of the various austenitic product forms used in the Safety Standards are also representative for the material of the austenitic cladding. Fatigue tests were performed to determine and assure experimentally a fatigue life curve of the cladding material of pressurized components and to compare the results with the database of the stabilized austenitic stainless steels used in German nuclear power plants.

To carry out fatigue tests with cladding material, flat specimens specially adapted to the geometric conditions were prepared out of the austenitic cladding of a reactor pressure vessel which had been manufactured according to nuclear specifications. Due to the small thickness of the cladding (max. 7 mm), specially designed flat specimens (thickness of 2 mm) had to be used instead of the normal cylindrical fatigue specimens. To check a possible influence of this specimen geometry, flat specimens were prepared in advance from an austenitic pipe of known fatigue behaviour and tested at two different strain amplitudes. A comparison of the results with the fatigue life curve determined with conventional cylindrical specimens showed no influence at the higher strain amplitudes, Fig. 9. At the lower strain amplitude, a small influence of the specimen geometry becomes apparent. The flat specimens show a slightly lower fatigue life than cylindrical specimens.

24 strain-controlled uniaxial fatigue tests under fully reversed conditions (R=-1) were carried out at RT with flat specimen, Fig. 10.

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06 1,0E+08 1,0E+10

Str

ain

Am

pli

tud

e

ε ε ε ε a

/ %

Load Cycles N / -

Smooth cylindrical specimen

MPA Mean Data

Flat specimen base material

Fig. 9 : Fatigue life of stabilized austenitic stainless steels at RT in air, comparison between cylindrical and flat specimen

163

At the highest strain amplitude (0,8 %), a slightly higher fatigue life was detected, whereas at the lowest strain amplitude (0,22 %), the fatigue life is slightly lower than that of the austenitic base material. Altogether, the fatigue life data with flat specimen made of the material of austenitic cladding fit well into the scatter band of the fatigue database of the austenitic base materials tested with cylindrical specimens.

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06 1,0E+08 1,0E+10

Str

ain

Am

pli

tud

e

ε ε ε ε a

/ %

Load Cycles N / -

Smooth cylindrical specimen

MPA Mean Data

Flat specimen base material

Flat specimen cladding

Fig. 10 : Fatigue life of stabilized austenitic stainless steel and austenitic cladding at RT in air

3.2.3 Ferritic material

More than 80 fatigue data available at MPA at RT for the low alloy ferritic steel 20MnMoNi5-5 are shown in Fig. 11. It includes is also the best-fit curve representing the low alloy steel data of [9], by

).(ln..)Nln( a 151080814496 −ε−= (7)

The MPA data are represented by this equation developed at ANL [9].

0,01

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Load Cycles / N

Best-Fit Curve RT

Ferrit RT

Fig. 11 : Fatigue life of low alloy steel 20MnMoNi5-5 at RT in air

164

More than 60 fatigue data available at MPA at temperatures between 240 oC and 350 oC for the low alloy ferritic steels 20MnMoNi5-5 and 22NiMoCr3-7are shown in Fig. 12. It includes is also the best-fit curve representing the low alloy steel data of ANL [9] for a temperature of 300 oC according to equation (3).

0,01

0,10

1,00

10,00

1,0E+00 1,0E+02 1,0E+04 1,0E+06

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Load Cycles / N

Best-Fit Curve, T=300°C

Ferrit Elev. Temp.

Fig. 12 : Fatigue life of low alloy steel2 20MnMoNi5-5 and 22NiMoCr3-7at temperatures between 240 oC and 350 oC in air

3.3 LCF tests in BWR environment

The investigations of environmental effects on fatigue life of stainless steels performed in the USA, e.g. [9], and in Japan, e.g. [12], are predominantly based on unstabilized austenitic stainless steels. There is a need to investigate whether the fatigue behaviour of the niobium or titanium stabilized austenitic stainless steels used in German nuclear power plants in contact with oxygenated high temperature (HT) water of boiling water reactor (BWR) coolant can be predicted by the curves and methods developed in the USA and in Japan. For the experiments a test facility was available including a miniature autoclave positioned on the specimen shoulders.

3.3.1 Austenitic material

The specimens are machined from 2 nuclear grade pipe materials of the stabilized austenitic stainless steels X10CrNiNb18-9 and X10CrNiTi18-9 (se chapter 3.1). The specimens of steel X10CrNiTi18-9 are additionally tested in a sensitized material state after an annealing treatment at 620 oC for 8 hours.

Low cycle fatigue (LCF) tests are conducted in simulated BWR HT-water environment at 240°C, the oxygen content of the water is adjusted to 400 ppb, [18] und [19]. In addition to high purity water, few tests are performed at an increased conductivity of 0.8 µS/cm by dosing sulphate (90 ppb SO4). The flow velocity is quasi stagnant (0.004 m/s). All experiments in water were preceded by a soaking period of 100 h at test conditions. The LCF tests were conducted with strain amplitudes of 0.3 %, 0.6 %, 0.9 % and 1.2 % and an uniform strain rate of 0.01 %/s.

The fatigue life in simulated BWR environment with high purity is conservatively covered by the ANL mean water curve [9], Fig. 13. The influence of the simulated BWR environment (Nair/Nwater) tends to decrease with increasing strain amplitude. Included in Fig. 13 is the

165

effective Fen value based on the test parameters and MPA test results in air. All Fen values are well below the value of 3.55 calculated by the ANL equation.

102

103

104

105

0.1

1

2.5

1.91.6

Fen

2.9

2.1

1.61.9

1.7

Austenitic SS

X10CrNiNb18-9

X6CrNiNb18-10

X10CrNiTi18-9

high purity water 240°C

ANL: Fen

= 3.55

* basis for Design Curve up to ASME Code 2008 ASME Design 2010

5

0.5

Reg.Guide 1.207 (N/Fen

)

ANL air 2007

KTA 3201.2

Reg.Guide 1.207

ANL water 2007

ASME air*

Str

ain

Am

plitu

de

εε εε a

[%

]

Cycles N

.εεεε = 0.01 %/s

E 240 °C = 183 000 MPa

Fig. 13 : Fatigue life of stabilized austenitic stainless steels in high purity HT-water compared with ANL mean water curve [9] and fatigue design curves

102

103

104

105

0.1

1

0.5

5


)

2.3

2.2

3.1 2.7 2.5

1.9

X10CrNiTi18-9

as delivered air

as delivered water

sensitised water

as delivered sulphate

E 240 °C = 183 000 MPa

ANL water 2007

Str

ain

Am

plitu

de

εε εε a

[%

]

Cycles N

KTA 3201.2

ASME Design 2010

ANL air 2007

ASME air

Reg.Guide 1.207

.

.F

en 1.6

air: 240 °C ε ε ε ε = 0.1 %/s

water: 240 °C εεεε = 0.01 %/s

ANL: Fen

= 3.55

Fig. 14 : Fatigue life of Ti-stabilized austenitic stainless steel (as delivered and sensitized material state) in high purity and sulphate HT-water compared with ANL mean water curve [9] and fatigue design curves

Under the given test parameters the sensitized material state of material X10CrNiTi18-9 as well as dosing a sulphate content of 90 ppb result only in a minor reduction of fatigue life,

166

Fig. 14. Included in Fig. 14 are again the effective Fen values based on the test parameters. Again all Fen values are below value of 3.55 calculated by the ANL equation.

3.3.2 Ferritic material

Fatigue experiments up to more than 106 load cycles were performed in high purity water as well as in sulphate (90 ppb SO4) and chloride (50 ppb Cl) respectively at a test temperature of 240 oC and an oxygen content of 400 ppb (simulated BWR coolant) [20]. The specimens are machined from a reactor pressure vessel (RPV) shell of low alloy steel 22NiMoCr3-7.

In the LCF tests strain amplitudes of 0.3 %, 0.5 % and 0.9 % were applied at strain rates of both 0.1 %/s and 0.01 %/s. The HCF experiments were conducted as stress-controlled cyclic tests (R=-1) at a loading frequency of 1 Hz (stress amplitude 325 MPa, nominal strain amplitude 0.166 %).

In high purity water at the “fast” strain rate of 0.1 %/s the decrease in fatigue life is in the range of the ANL prediction (mean curve), Fig. 15. On the contrary, the “slow” strain rate of 0.01 %/s causes a drop of fatigue life in high purity water, Fig. 15. The effect of sulphate and chloride on the fatigue life is hardly more pronounced than of high purity water whereas the corrosive effect of the medium, in particular of the sulphate, increases with decreasing strain rate, Fig. 16.

101

102

103

104

105

106

107

0.1

1

10

Fen

2.4 1.7

3.5 2.1

6.4 2.6

Str

ain

Am

plitu

de

εε εε a

[%

]


)

0.66 %/s


)

0.01 %/s

ANL water 2007

0.66 %/s

0.1 %/s LCF strain controlled

0.01 %/s LCF strain controlled

1 Hz HCF stress controlled

water (high purity) 240°C

Fen ANL = 2.11 / 2.57 / 3.27

ANL water 2007

0.01 %/s


)

0.1 %/s

ANL water 2007

0.1 %/s

ASME mean air

ASME Design 2007

KTA 3201.2

LAS 22NiMoCr3-7

Cycles N

Fig. 15 : Fatigue life of low alloy steel 22NiMoCr3-7 in high purity HT-water compared with ANL mean water curve [9] and fatigue design curves

Fig. 17 shows the results in high purity and sulphate containing HT-water at the slow strain rate of 0.01 %. Obviously the sulphate causes a concentration of the fatigue life results at the lower bound of the water results.

3.4 Crack growth tests in BWR environment

Crack growth tests were performed with fracture mechanics specimen C(T)25, C(T)50 and C(T)100 in high purity water with an oxygen content of 400 ppb as well as in chloride containing water (50 ppb Cl) at a test temperature of 288 oC (simulated BWR coolant) [21]. The specimens are machined from a reactor pressure vessel (RPV) shell of low alloy steel 22NiMoCr3-7.

167

All specimen were pre-cycled in high purity water to create a detectable crack size, typically in 120 sec time period (100 sec load increase, 20 sec load decrease), in few cases in 12 sec time period.

101

102

103

104

105

106

107

0.1

1

10

4.8 3.3 2.6

2.1 2.1 2.2

Fen

2.1 1.7 1.6

LAS 22NiMoCr3-7

water 0.1 %/s LCF strain cont.

water 1 Hz HCF stress cont.

sulphate (90 ppb) 0.1 %/s LCF strain cont.

sulphate (90 ppb) 1 Hz HCF stress cont.

chloride (50 ppb) 0.1 %/s LCF strain cont.

240 °C Fen ANL = 2.11 / 2.57

Reg.Guide 1.207 (N/Fen)0.66 %/s

ASME Design 2007

KTA 3201.2

ANL water 2007

0.66 %/s

Str

ain

Am

plitu

de

εε εε a

[%

]

Reg.Guide 1.207 (N/Fen)

0.1 %/s

ANL water 2007

0.1 %/s

ASME mean air

Cycles N

Fig. 16 : Fatigue life of low alloy steel 22NiMoCr3-7 in high purity, sulphate and chloride HT-water compared with ANL mean water curve [9] and fatigue design curves

101

102

103

104

105

106

0.1

1

10

12.3 6.4

4.8 3.5

Fen

3.2 2.4

ASME Design 2007

KTA 3201.2

Str

ain

Am

plitu

de

εε εε a [%

] ANL water 2007

0.01 %/s


)

0.01 %/s

ASME mean air

water sulphate 90 ppb

0.01 %/s 240°C

Fen ANL

= 3.27

LAS 22NiMoCr3-7

Cycles N

Fig. 17 : Fatigue life of low alloy steel 22NiMoCr3-7 in high purity and sulphate HT-water at slow strain rate compared with ANL mean water curve [9] and fatigue design curves

The data of all measured crack growth rates for the specimen geometries tested and the applied R ratios are between the ASME air and water curves, Fig. 18 and Fig. 19, with the

168

exception of 2 C(T)25 specimen with R=0.7 located on the ASME water curve. Thus the measured crack growth rates can be rated as moderate.

A reduction of the 120 sec time period to 12 sec increases the crack growth rate up to a factor of 2, i.e. a reduced velocity in load increase under high purity water conditions is not connected with an increase in crack growth. The cyclic crack growth rates increase as expected with higher values of ∆∆∆∆K.

10-2

10-1

100

101

Cra

ck G

r ow

th R

ate

da/d

N [

µm

/cycle

]

1 10 100

∆∆∆∆ KI [MPa√√√√m]

C(T)25 120 s/cycle

C(T)50 120 s/cycle

C(T)50 12 s/cycle

C(T)100 120 s/cycle

air

ASME XIWasserR0.7

wat

er

Fig. 18 : Cyclic crack growth in high purity water at 288 oC with R ratio 0.7

10-2

10-1

100

101

Cra

ck G

r ow

th R

ate

da/d

N [

µm

/cycle

]

1 10 100

∆∆∆∆ KI [MPa√√√√m]

C(T)25 120 s/cycle

C(T)50 120 s/cycle

C(T)100 120 s/cycle

ASME XI WasserR0.1

wat

erair

Fig. 19 : Cyclic crack growth in high purity water at 288 oC with R ratio 0.1

169

To investigate the influence of specimen size and R ratio (KI min/KI max) on the crack growth behaviour tests were performed with the three specimen sizes and the loading sequence as shown in Fig. 20 and Fig. 21. Subsequent to the cyclic loading 9 partial unloading/reloadings (PUR) each followed by a constant load phase of 50 h at a KI value of 80 MPa√m were applied.

0

20

40

60

80

100

120

Str

es

s I

nt e

ns

ity

K

I [

MP

a√√ √√m

]

-100 0 100 200 300 400 500 600

Time [h]


cyclic 9 PUR

R = 0.7

C(T)25 / C(T)50 / C(T)100

50 h

co

nd

itio

nin

g p

ha

se

Fig. 20 : Loading sequence (PUR) for R ratio 0.7 (schematic)

0

20

40

60

80

100

120

Str

es

s I

nt e

ns

ity

KI

[M

Pa

√√ √√m

]

-100 0 100 200 300 400 500 600

Time [h]

cyclicR = 0.1

9 PUR

C(T)25 / C(T)50 / C(T)100

50 h


co

nd

itio

nin

g p

ha

se

R = 0.1

Fig. 21 : Loading sequence (PUR) for R ratio 0.1 (schematic)

During PUR and the constant load phases at no test pronounced crack growth was observed. Crack Growth measurement by DCPD under high purity water conditions is shown for C(T)50 with R=0.7 in Fig. 22 and for C(T)100 with R=0.1 in Fig. 23. The crack growth rates caused by PUR are in the same range as per load cycle during cyclic loading. For the C(T)50 specimen crack growth during cyclic loading accounts 0.26 µm/cycle, 0.11 µm for PUR and the C(T)100 specimen crack growth during cyclic loading 3.48 µm/cycle, 5.78 µm for PUR.

As a result for the high KI max value of 80 MPa√m under high purity water conditions no significant influence of specimen size at cyclic and transient loading to crack growth behaviour was observed. The fractographical analysis showed no indications of stress corrosion cracking (SCC).

170

To investigate the influence of chloride C(T)25 and C(T)50 specimen with KI max values of 40 and 55 MPa√m were tested. Loading scheme see Fig. 24.

0

100

200

300

400

500

600

700

800

900

1000

Cra

ck G

row

th

[µ

m]

-100 -50 0 50 100 150 200 250 300 350 400 450 500 550

Time [h]

2182cycles

end ofcycling

PUR1

PUR2

PUR3

PUR4

PUR5

PUR6

PUR7

PUR8

PUR9

end506 h

C(T)50 80 MPa√√√√m9 unloadings/reloadings R0.7high purity water 288°C

fracture surface: ∆∆∆∆a = 565 µm

partial unloading/reloading0.11 µm/PUR (average)

cycl

ing:

0.2

6 µ

m/c

ycle

Fig. 22 : DCPD crack growth measurement, C(T)50 specimen with R ratio 0.7

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

Cra

ck G

row

th

[µ

m]

-50 0 50 100 150 200 250 300 350 400 450 500 550

Time [h]

C(T)100 80 MPa√√√√m9 unloadings/reloadings R0.1high purity water 288°C


partial unloading/reloading5.78 µm/PUR (average)

250cycles

cycl

ing:

3.4

8 µ

m/c

ycle

end ofcycling

PUR1

PUR2

PUR3

PUR4

PUR5

PUR6

PUR7

PUR8

PUR9

end507 h

Fig. 23 : DCPD crack growth measurement, C(T)100 specimen with R ratio 0.1

After starting the chloride transient of 50 ppb during the first phase of constant load no crack growth was detected, i.e. no SCC occurred in chloride containing water under static load conditions. In contrary, severe SCC crack growth was caused by the load transients, which can be divided into a first phase (phase 1) with a high crack growth rate and a second one (phase 2) with a decreasing crack growth rate. The total SCC crack propagation during these experiments was in the range between 270 and 1150 µm, example see Fig. 25.

The crack growth behaviour caused by the load transients (PUR) at a chloride content of 50 ppb was similar for the stress intensities of 40 and 55 MPa√m. Compared to the BWR VIP-60 disposition lines (DL) [22] the crack growth rates during phase 1 are between the low sulphur line and the high sulphur line, whereas for 55 MPa√m most of the crack growth rates of phase 2 are below DL2 and the crack growth rates of phase 2 are between DL2 and DL1, Fig. 26 .

171

0

20

40

60

80

100

Str

ess In

t en

sit

y K

I [

MP

a√√ √√m

]

-100 0 100 200 300 400 500 600

Time [h]

high purity water / chloride transient 288°C

cyclicR = 0.7

PUR 1R = 0.7

C(T)25 / C(T)50

100 hc

on

dit

ion

ing

ph

as

e

PUR 2R = 0.4

PUR 3R = 0.1

50 ppb chloridewater water

Fig. 24 : Loading sequence (PUR) under chloride transient loading (50 ppb chloride)

0

250

500

750

1000

1250

1500

1750

2000

2250

Cra

ck

Gro

wth

[

µm

]

-100 -50 0 50 100 150 200 250 300 350 400 450 500 550 600

Time [h]

C(T)25 55 MPa √√√√m3 PUR R0.7, R0.4, R0.1high purity water / chloride 288 °C

end ofcycling

startchloride

PUR1

PUR3

PUR2

stopchloride

end593.3 h

50 ppb chloride waterwater


cycling0.142

µm/cycle

cycling0.705

µm/cycle

0.035µm/h

0µm/h

0.683µm/h

0.387µm/h

0.463µm/h

0µm/h

1587cycles

13.61µm/h

21.79µm/h

10.84µm/h

phase 1

phase 2

Fig. 25 : DCPD crack growth measurement under chloride transient loading

4 Fatigue mean data and design curves

Based on the evaluations of fatigue data at ANL [9] (NUREG/CR-6090) new fatigue design curves are implemented in ASME-Code Section III (since ed. 2008 and later) [1] for austenitic stainless steels, Fig. 6. To develop a fatigue design curve factors on stress respectively strain and cycles shall be applied to the best-fit curve as already mentioned (for austenitic stainless steel factors Sσσσσ=2 and SN=12 are implemented).

Fatigue test data at RT and at elevated temperatures for the German stabilized austenitic stainless steels X10CrNiNb18-9, X6CrNiNb18-10 and X10CrNiTi18-9 are represented by the best-fit curves according to eq. (7), Fig. 7, respectively eq. (8), Fig. 8.

172

To define the factor on stress useful values for surface, size and mean stress (verified by tests and used in conventional vessel and piping standards) as well as the data scatter for RT date result in a value of Sσσσσ=1.88, Fig. 27. In this way in connection with SN=12 especially in the range of endurance limit this new fatigue design curve for RT based on the best-fit curve eq. (7) is approximately 1,4 above the ASME-Code design curve.

20 30 40 50 60 70 80 90 10010

-12

10-11

10-10

10-9

10-8

10-7

10-6

C(T)25 Phase 1

C(T)25 Phase 2

C(T)50 Phase 1

C(T)50 Phase 2

288 °C 0.4 ppm O2 50 ppb Chlorid

crack growth caused by load transients

Phase 2

DL 2

DL 1

BWRVIP-60 Disposition Lines (DL)

low sulphur material

high sulphur material

Cra

ck

Gro

wth

Rate

d

a/d

t [

m/s

]

Stress Intensity KI [MPa√√√√m]

Phase 1

Fig. 26 : Crack growth under chloride transient loading compared to BWRVIP-60 Disposition Lines [22]

0,01

0,10

1,00

10,00

1,0E+01 1,0E+03 1,0E+05 1,0E+07 1,0E+09 1,0E+11

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Load Cycles / N

Best-Fit Curve, Room Temperature

ASME Design Curve, ed. 2010

Design Curve (SN=12, S_sig=1.88)

S_N

S_sig

Fig. 27 : Design curve at room temperature

173

To define the factor on stress in same way like for RT a value of Sσσσσ=1.79 shall be used. In this way in connection with SN=12 the new fatigue design curve based on the best-fit curve eq. (8) shows nearly the same behavior like the ASME-Code design curve, Fig. 28.

0,01

0,10

1,00

10,00

1,0E+01 1,0E+03 1,0E+05 1,0E+07 1,0E+09 1,0E+11

Str

ain

Am

pli

tud

e ε

ε ε ε a

/ %

Load Cycles / N

Best-Fit Curve, Elevated Temperatures

ASME Design Curve, ed. 2010

Design Curve (SN=12, S_sig=1,79)

S_N

S_sig

Fig. 28 : Design curve at elevated temperatures

References

[1] ASME Boiler and Pressure Vessel Code Section III, Rules for Construction of Nuclear Facility Components, Division 1 – Subsection NB Class 1 Components, Rules for Construction of Nuclear Facility Components, Materials, Division 1 – Appendices, Appendix I – Design Fatigue Curves, The American Society of Mechanical Engineers, New York, 2010 Edition ), 2011 Addenda (July 1, 2011)

[2] French Design and Construction Rules for Mechanical Components of PWR Nuclear Islands (RCC-M). AFCEN - Association Française pour la Construction des Ensembles Nucléaires, Paris

[3] Safety Standards of the Nuclear Safety Standards Commission (KTA). KTA Rules 3201 and 3211, Carl Heymanns Verlag KG, Cologne, latest edition

[4] Langer B.F., "Design of Pressure Vessels for Low-Cycle Fatigue", Journal of Basic Engineering, Vol. 84, 3 (1962)

[5] Jaske C.E., W.J. O'Donnell, "Fatigue Design Criteria for Pressure Vessel Alloys", Journal of Pressure Vessel Technology (1977)

[6] Diercks D.R., "Development of Fatigue Design Curves for Pressure Vessel Alloys using a Modified Langer Equation", Journal of Pressure Vessel Technology, Vol. 101 (1979)

[7] Criteria of Section III of the ASME Boiler and Pressure Vessel Code for Nuclear Vessels, ASME 1964, Library of Congress Catalog Card Number: 56-3934, and

174

Criteria of the ASME Boiler and Pressure Vessel Code for design by analysis in Section III and VIII, Division 2, ASME 1969, Library of Congress Catalog Card Number: 56-3934

[8] Fatigue strength reduction and stress concentration factors for welds in pressure vessels and piping, Welding Research Council Bulletin, WRC 432, June 1998

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

[10] Chopra O.K., Effects of LWR Coolant Environments on Fatigue Design Curves of Carbon and Low Alloy Steels, NUREG/CR-6583, ANL-97/18, March 1998

[11] Rosinski S., Materials Reliability Program (MRP), Evaluation of Fatigue Data Including Reactor Water Environmental Effects (MPR-49), EPRI Technical report 1003079, Dec. 2001

[12] Higuchi M., Japanese Program Overview, 3rd Intern. Conference on Fatigue of Reactor Components, Seville, Spain, October 3 – 6, 2004

[13] O’Donnell, W.J., P.T. O’Donnell, W.J. O’Donnell, Proposed new fatigue design curves for carbon and low-alloy steels in high temperature water, Journal of pressure vessel technology, Vol. 131, April 2009,

[14] Nomura, Y., K. Tsutsumi, T. Inoue, S Asada, T, Nakamura, Optimization of environmental fatigue evaluation (step 2), ASME PVP2009-77115

[15] Chopra O.K, Y. Garud, Update of NUREG/CR-6909 methodology for environmentally assisted fatigue (EAF) - revised Fen expressions, ASME Code meetings, Section III, Subgroup on fatigue strength, Nashville, TN, May 15, 2012

[16] Guidelines for Evaluating Fatigue Analyses Incorporating the Life Reduction of Metal Components due to the Effects of the Light-Water Reactor Environment for New Reactors, Regulatory Guide 1.207, U.S. Nuclear Regulatory Commission, March 2007

[17] Keisler, J., O.K.Chopra, W.J. Shack, Fatigue strain-life behavior of carbon and low-alloy steels, austenitic stainless steels, and alloy 600 in LWR environments, NUREG/CR-6335, ANL-95/15, August 1995

[18] Weißenberg, T., Zentrale Untersuchung und Auswertung von Herstellungsfehlern und Betriebsschäden im Hinblick auf druckführende Anlagenteile von Kernkraftwerken, Arbeitspaket 3, Einfluss des Reaktorkühlmediums auf das Ermüdungsverhalten austenitischer Rohrleitungen, BMU-Vorhaben SR 2501, MPA Universität Stuttgart, November 2007

[19] Weißenberg, T., Zentrale Untersuchung und Auswertung von Herstellungsfehlern und Betriebsschäden im Hinblick auf druckführende Anlagenteile von Kernkraftwerken, Arbeitspaket 3.1, Untersuchung des Einflusses von Reaktorkühlmedium auf das Ermüdungsverhalten austenitischer CrNi-Stähle, BMU-Vorhaben SR 08 01312, MPA Universität Stuttgart, Juni 2011

[20] W Weißenberg, T., Ermüdungsverhalten ferritischer Druckbehälter- und Rohrleitungsstähle in sauerstoffhaltigem Hochtemperaturwasser, Abschlussbericht, BMWi Reaktorsicherheitsforschung - Kennzeichen 1501309, MPA Universität Stuttgart, September 2009

[21] Weißenberg, T., Rissverhalten ferritischer Druckbehälterstähle in sauerstoffhaltigem Hochtemperaturwasser bei transienten Vorgängen, Risskorrosion Phase 1: Riss-initiierung und Risswachstum, Teilbericht A: Experimentelle Risskorrosionsunter-suchungen, Abschlussbericht, Forschungsvorhaben 1501319, MPA Universität Stuttgart, 11/2010

175

[22] BWR VIP-60, BWR Vessel and Internals Project, "Evaluation of Stress Corrosion Crack Growth in Low Alloy Steel Vessel Materials in BWR Environment (BWRVIP-60)" EPRI Technical Report TR 108709 (1999)

176

fatigue behaviour and crack growth of ferritic steel under ......fatigue tests were performed in the...

Documents