High-temperature low cycle fatigue, creep–fatigue and thermomechanical fatigue of steels and their welds

ARTICLE IN PRESS

International Journal of Mechanical Sciences 48 (2006) 160–175 www.elsevier.com/locate/ijmecsci

High-temperature low cycle fatigue, creep–fatigue and thermomechanical fatigue of steels and their welds S.L. Mannan, M. Valsan Metallurgy and Materials Group, Indira Gandhi Centre for Atomic Research, Kalpakkam 603 102, Tamil Nadu, India Received 25 February 2005 Available online 12 October 2005

Abstract High-temperature low cycle fatigue (LCF) is influenced by various time-dependent processes such as creep, oxidation, phase transformations and dynamic strain ageing (DSA) depending on test conditions of strain rate and temperature. In this paper, the detrimental effects of DSA and oxidation in high-temperature LCF are discussed with reference to extensive studies on 316L(N) stainless steel and modified 9Cr–1Mo steel. DSA has been found to enhance the stress response and reduce ductility. It localizes fatigue deformation, enhances fatigue cracking and reduces fatigue life. High-temperature oxidation accelerates transgranular and intergranular fatigue cracking in modified 9Cr–1Mo steel and during long hold time tests in austenitic stainless steel. In welds, microstructural features such as presence of course grains in the HAZ and formation of brittle phases due to transformation of d ferrite during testing influence crack initiation and propagation and fatigue life. Thermomechanical fatigue (TMF) studies are suggested as more closer to the actual service conditions. In 316L(N) stainless steel, TMF lives under out-of-phase cycling are found to be lower than those under in-phase conditions in the low-temperature regimes, while the converse holds good when the upper temperature encompassed the creep-dominant regime. r 2005 Elsevier Ltd. All rights reserved. Keywords: Low cycle fatigue; Thermomechanical fatigue; 316L(N) steel; Modified 9Cr–1Mo steel; Welds

1. Introduction Low cycle fatigue (LCF) is an important consideration in the design of high-temperature systems subjected to thermal transients. The systems that experience thermal transients include aircraft gas turbines, nuclear reactor vessels, heat exchangers, steam turbines and other power plant components. LCF resulting from thermal transients occurs essentially under strain controlled conditions, since the surface region is constrained by the bulk of the component. In thick components the major compressive strain is introduced by the thermal transient during start up, with additional compressive and or tensile strains during load cycling and shut down. On-load periods at elevated temperatures in between transients introduce timeCorresponding author. Metallurgy and Materials Group, IGCAR, Kalpakkam 603 102, Tamil Nadu, India. Tel.: +91 4114 280 222; fax: +91 4114 280 081. E-mail address: [email protected] (S.L. Mannan).

0020-7403/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmecsci.2005.08.004

dependent effects. Temperature time transients which could be experienced by components in a fast reactor are the result of reactor trip (down shock) or secondary circuit failure (up shock) [1]. Thermo mechanical fatigue (TMF) with simultaneous mechanical and thermal cycling is more close to service situations. Traditionally, isothermal LCF tests have been used to assess the performance of materials subjected to thermal transients. Thus, the component behaviour is studied using mechanical strain cycling under isothermal testing conditions. The slow start up/shut down cycle is replaced by a symmetrical and continuous fatigue cycle of equal strain rates in tension and compression with a hold period at constant peak strain to simulate the on-load period. At high temperatures the fatigue deformation and life are influenced by several time-dependent mechanisms such as dynamic strain ageing (DSA), oxidation, creep and phase transformations. These damage processes, which are strong functions of temperature and strain rate, are

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

illustrated with examples from extensive studies conducted on 316L(N) stainless steel and their welds and modified 9Cr–1Mo ferritic steel [2–12] which are the currently favoured structural materials for the primary and secondary sides respectively of the liquid metal-cooled fast reactor. In this paper, high-temperature fatigue and creep–fatigue properties of 316L(N) stainless steel, its welds and modified 9Cr–1Mo steel are discussed. This is followed by a brief discussion of TMF properties of 316L(N) steel. 1.1. 316L(N) stainless steel and welds Nitrogen modified 316L(N) austenitic stainless steel is used in nuclear power plants for the construction of reactor vessel, piping and heat exchangers. This alloy has emerged as a viable alternative to AISI 316 and its modified grades, with enhanced high-temperature mechanical properties and lesser susceptibility to sensitization and associated intergranular corrosion. Evaluation of elevated temperature LCF behaviour of 316L(N) stainless steel has received much attention in the recent years [2,13–19]. In these studies, nitrogen addition has been reported to be beneficial, and the LCF life has been found to improve at both ambient temperature and 873 K. Welds are the weak links in structures. Most of the service failures are found to occur either in the HAZ or in the weld metal, which are more frequently associated with the presence of defects or microstructural inhomogeneities compared to the base metal. In austenitic stainless steel welds, the d ferrite introduced to reduce their tendency to hot cracking and micro fissuring, transforms to a hard and brittle phase known as s phase, when these materials are exposed to elevated temperatures (773–1173 K) for extended periods of time, leading to low ductility creep ruptures when sufficiently high stresses are applied at elevated temperatures. The allowable number of fatigue cycles in welds is onehalf the value permitted for parent material as per the ASME Boiler and Pressure Vessel Code [20] and a factor of 1.25 on fatigue strain is applied as per the RCC-MR design code [21].

161

also possesses better monotonic tensile and creep strengths at elevated temperatures compared with the plain 9Cr–1Mo steel. The alloy also exhibits good weldability and microstructural stability over very long periods of exposure to high-temperature service conditions. The addition of Nb improves the properties by promoting nucleation of finely distributed M23 C6 carbides and by aiding grain size refinement, whereas V enters the carbide particles and retards their growth. It must be mentioned that in this alloy, strength in normalized and tempered condition is derived from carbides like NbC, VC and M23 C6 on sub-boundaries and from the tempered martensitic laths with high dislocation densities. In addition, V and Nb could also form fine precipitates of nitrides/ carbonitrides within the ferrite matrix contributing to further strengthening [23–26]. Mo is a solid solution strengthener and is considered a retardant for dislocation recovery/recrystallization [27]. LCF behaviour of modified 9Cr–1Mo ferritic steel has been reported earlier under normalized and tempered [25,28–32,9,11] and thermally aged conditions [27,33,34]. Further, detailed investigations have been carried out to evaluate the creep–fatigue interaction behaviour of the alloy [27,28,33–36]. Ebi and McEvily [25] showed that the hot forged alloy exhibits inferior fatigue properties at 811 K as compared to the fine-grained, hot-rolled material. It was concluded that the coarse grain size adversely affects the fatigue crack initiation stage but had little effect on the crack propagation. Prolonged ageing of the alloy at elevated temperatures prior to testing was found to reduce the LCF and creep–fatigue interaction lives [27,34]. Ageing resulted in the formation of Laves phase with associated reduction in the toughness and LCF life of the alloy [27,34]. Studies on the influence of strain hold position on the creep–fatigue life indicated that the hold in compression peak strain was more deleterious than that in tension. This was attributed to the detrimental effect of oxide behaviour in compression hold [36]. Life was observed to decrease with increase in the dwell time up to 1 h under tension hold beyond which there was an apparent saturation [27]. During long time creep tests, carbide coarsening and coalescence at grain boundaries were found to introduce intergranular damage in the alloy [37].

1.2. Modified 9Cr– 1Mo steel 2. Experimental Modified 9Cr–1Mo ferritic steel (with alloying additions of niobium and vanadium and controlled amount of nitrogen) is extensively used as a structural material at elevated temperatures up to 873 K in fossil-fired power plants, petrochemical industries and as a material for tubing in the reheater and superheater portions and as thick-section tube sheet material in the steam generators of liquid metal cooled fast breeder reactor [22]. High thermal conductivity and low thermal expansion coefficient coupled with enhanced resistance to stress corrosion cracking in steam–water systems are important considerations in the selection of this steel for these applications. The material

The 316L(N) base metal in the mill-annealed condition was solution treated at 1373 K for 1 h, followed by water quenching prior to machining the LCF specimens. Weld metal specimens were machined from weld pads prepared by shielded metal arc welding process using 316 and 316(N) electrodes. X-ray radiography was used for assessing the soundness of the welds followed by d ferrite measurements using a magne-gauge. Modified 9Cr–1Mo steel was obtained in the form of hot-forged rods of 70 mm diameter. The normalizing treatment of this steel was carried out at 1313 K for 1 h plus air cooling and tempering was done at

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

162

1033 K for 1 h plus air cooling which resulted in a tempered martensitic structure. 2.1. Low cycle fatigue and thermomechanical fatigue testing Fully reversed total axial strain controlled LCF tests were conducted at 773, 823 and 873 K in air on the 316L(N) base metal, 316 weld metal and 316(N) weld metal and modified 9Cr–1Mo specimens using a servo hydraulic machine equipped with a radiant heating facility. Tests were carried out with total strain amplitudes in the range
0:25% to
1:0% with the strain rates varying from

±0.05 ∅8.38

+0.00 −0.05 ∅15.75

±0.05 ∅8.65

139.8 25.4R

32

38

3  105 to 3  102 s1 . Creep–fatigue interaction experiments were conducted by introducing tension/compression hold in the range 1–90 min at 873 K. In TMF tests, both inphase (wherein the maximum temperature and peak tensile strain coincide) and out-of-phase (maximum temperature coincides with peak compressive strain) experiments were carried out on 316L(N) austenitic stainless steel on hollow tubular samples (Fig. 1) using different mechanical strain amplitudes in the range,
0:25% to
0:6% at a constant strain rate of 6:4  105 s1 . A schematic representation of the temperature, stress and strain relationships in the IP and OP tests is shown in Fig. 2. The temperature cycles employed consisted of an identical temperature range with increasing peak temperature (300–550, 350–600, 400–650 1C).

±0.05 ∅11.43

2.2. Metallography

0.8 × 45°

44.5

The tested samples were sectioned parallel to the loading direction, polished, etched and examined under an optical microscope. The 316L(N) base metal was etched using 70%

50.8

Fig. 1. Geometry of TMF test specimen.

0.02

400 εm

εth

εt

t

ε

0.016

0.012 Stress (MPa)

Strain (mm /mm)

200

εth

0.008

εm

0.004

0 -0.01

-0.005

0

0.005

0.01

0.015

0.02

0.025

Strain (mm/mm) -200

0 63000

(a)

63500

64500

64000

-400

-0.004

(b)

Time (s)

400

m

ε

0.014 0.012

ε

t

ε

th

εt

200

0.008 0.006 εm

0.004

Stress (MPa)

Strain (mm /mm)

0.01

th

ε

0 0.005

0

0.005

0.01

0.015

0.02

Strain (mm/mm)

0.002 -200 0 300

500

700

900

1100 1300 1500 1700

-0.002 -400

-0.004

(c)

Time (s)

(d)

Fig. 2. (a) and (b). Waveform employed and hysteresis obtained in IP TMF tests (
th : thermal strain,
m : mechanical strain;
t : total strain), (c) and (d). Waveform employed and hysteresis obtained in OP TMF tests (
th : thermal strain;
m : mechanical strain;
t : total strain).

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

+1.0% +0.6% +0.4% +0.25%

TENSILE STRESS AMPLITUDE (MPa)

400

300

200

316L(N) BASE METAL TEMPERATURE :773 K STRAIN RATE: 3 x 10-3 s-1

100

10 0

(a)

perchloric acid and 80% methanol at 243 K with a d.c. voltage of 10 V. 3. Results and discussion 3.1. LCF of 316L(N) stainless steel: cyclic properties and role of DSA High-temperature LCF behaviour of 316L(N) stainless steel has been studied by several investigators [2–8]. The dependence of the peak tensile stress on the number of cycles and on total strain amplitude for the base metal at 773 K is depicted in Fig. 3(a). This material generally exhibits a very rapid strain hardening to a maximum stress 316L(N) SS 400 TENSILE STRESS AMPLITUDE, MPa

HNO3 while the etching of weld metal was done using a modified Murakami’s reagent (30 g of KOH, 30 g of K3 FeðCNÞ6 in 150 ml water) at 363 K for 30 s. Modified 9Cr–1Mo samples for the optical metallography were etched using Vilella’s reagent (1 gm of picric acid+5 ml conc. HCl+100 ml ethyl alcohol). Fractography of the failed specimens was carried out using a PSEM 501 scanning electron microscope and substructural changes were studied by Philips CM 12 transmission electron microscope. Samples for transmission electron microscopy (TEM) were obtained from thin slices cut at a distance of 3 mm away from the fracture surface. These samples were first mechanically polished down to 250 mm, and then electropolished in a solution containing 20%

163

STRAIN AMPLITUDE : + 0.6% TEMPERATURE

: 823 K

300

3 x 10-5 s-1

200

3 x 10-4 s-1 3 x 10-3 s-1 3 x 10-2 s-1

10 1 10 2 10 3 NUMBER OF CYCLES (N)

100 10 0

10 4

(b)

10 1 10 2 NUMBER OF CYCLES N

10 3

316L(N) SS ε : 3 x 10-3 s-1 ∆ε t / 2 : + 0.60%

316 LN 873 K -5 -1

360

3 × 10 s 5 4 3 2 1

280

80 −0.6

−0.4

−0.2

−80

0

0.2

0.4

0.6

STRAIN (%)

TENSILE STRESS AMPLITUDE, MPa

STRESS (MPa)

400

−280

(d)

298 573 673 773 873 923

200

100 10 0

−360

(c)

300

10 1

K K K K K K

10 2

10 3

NUMBER OF CYCLES N

Fig. 3. (a) Cyclic stress response—strain amplitude effect, (b) cyclic stress response—strain rate effect, (c) serrations on the stress—strain hysteresis loops and (d) cyclic stress response—temperature dependence.

ARTICLE IN PRESS 164

S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

followed by a nearly stable peak stress. It is observed that the life spent in the saturation stage is longer as the strain amplitude is lowered. The initial hardening observed in the base metal is mainly due to DSA caused by the interaction between dislocations and solute atoms. DSA was found to manifest in the cyclic stress response of 316L(N) in the temperature range 773–873 K, Fig. 3(b) (temperature 823 K), where a negative strain rate dependence of stress response is observed. The negative strain rate–stress response observed at these temperatures and strain rates was also reflected in a negative strain rate dependence of half-life stress [8]. Further, at all these temperatures, at strain rates o3  103 s1 , serrated flow was observed (Fig. 3(c)) and the cyclic stress response increased with increasing temperature in the range 573–873 K, Fig. 3(d). In the DSA regime, enhanced slip planarity and the degree of inhomogeneity of deformation are observed [8]. Further, the dislocation structure changed under the influence of increasing contribution from DSA. The solute dislocation interaction during DSA restricted cross slip of dislocations and increased the slip planarity. Dislocation structure changes from a cell structure at temperatures below 573 K, Fig. 4(a) to planar one between 573 and 873 K (DSA regime), Fig. 4(b) and back to a cell/subgrain microstructure beyond the DSA regime. Planar slip bands were observed in this material in the temperature and strain rate regime where DSA was predominant [16]. This exerted a profound influence on the fatigue life. In order to assess the influence of planar slip on cracking, the total length of intergranular and transgranular crack density was measured on the specimen surface in the DSA regime. Crack density (crack length per unit area) as a function of strain rate at 773 K is shown in Fig. 5(a) and the number of cracks (of different crack length) as a function of strain rate at 773 K is shown in Fig. 5(b). It is noticed that in the DSA regime, both transgranular and intergranular crack density increases with decrease in strain rate (Fig. 5(a)) and very long cracks are seen (a consequence of crack coalescence) at low strain rates. The stress concentration associated with the intersection of planar slip bands with the grain boundaries

could have contributed to the enhanced internal grain boundary cracks and reduced lives at low strain rates, Fig. 6. The continuous decrease in fatigue life with a decrease in strain rate and increase in temperature in the base metal is thus attributed to DSA under conditions where the effects of oxidation and creep are non-existent [8]. An increase in dislocation density with reduced strain rate was reported earlier in this material at 773 and 823 K [8]. It was pointed out that DSA would enhance the degree of inhomogeneity of deformation during LCF by the solute locking of slow moving dislocations between slip bands [38]. Presumably, the dislocation velocities inside the slip

Fig. 5. (a) Crack density as a function of strain rate and (b) crack length distribution as a function of strain rate.

Fig. 4. (a) Well developed cell structure at 573 K and (b) planar deformation bands at 873 K.

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

1600

0.010 PLASTIC STRAIN AMPLITUDE

316L(N) SS

NUMBER OF CYCLES TO FAILURE, Nf

∆ε t / 2 : + 0.6%

1200

800

165

STRAIN RATE: 3 x 10-3s-1 773 K 823 K 873 K

0.001 50 µm 1000 10000 NUMBER OF REVERSALS TO FAILURE Fig. 7. Fatigue life of modified 9Cr–1Mo at various temperatures.

400 773 K 823 K 873 K

0 10 -5

10 -4

10 -3

10 -2

Table 1 Half-life plastic strain amplitudes at different temperatures in modified 9Cr–1Mo steel

10 -1

STRAIN RATE / s-1 Fig. 6. Fatigue life as a function of strain rate.

bands were too high for dynamic ageing of mobile dislocations to take place and consequently DSA enhanced the partitioning of strains into separate regions characterized by high and low amplitudes of dislocation movement [39]. During DSA, slow moving dislocations become aged by the solute atmospheres and additional dislocations are generated to maintain the imposed deformation rate. This process caused an increase in the total dislocation density. The negative strain rate dependence of cyclic stress response over the temperature and strain rate range where DSA operates results from an increase in total dislocation density during deformation. The matrix was hardened during DSA, causing an increase in flow stress needed to impose the same total strain during successive cycles. 3.2. Role of oxidation and DSA in LCF of modified 9Cr– 1Mo steel Oxidation is observed to be an important damage mechanism in high-temperature LCF under testing conditions involving low strain rates and long hold times [9,10]. Modified 9Cr–1Mo steel exhibited a decrease in fatigue life with increasing temperature in the range 773–873 K. The effect of temperature on life was more pronounced at lower strain amplitudes ð
0:25%Þ, as seen in Fig. 7. Ingress of oxygen and oxidation of surface-connected grain boundaries and slip bands were observed at lower strain amplitudes [12]. Failure at 873 K was found to be strongly

Total strain amplitude, Plastic strain amplitude, D
p =2 (K) D
t =2 (%) 873 823 773 0.25 0.40 0.60 1.00

0.07 0.22 0.35 0.77

0.09 0.23 0.44 0.78

0.11 0.27 0.42 0.83

influenced by oxidation, particularly at lower strain amplitude of testing ð
0:25%Þ. The decrease in fatigue life with increase in temperature has been attributed to the combined effects of oxidation and DSA. The alloy displayed certain evidences for the occurrence of DSA in the temperature range 773–873 K. The half-life plastic strain amplitude decreased with increase in the test temperature, Table 1. Such negative temperature dependence of half-life plastic strain amplitude is a typical manifestation of DSA in LCF [11]. It must be pointed out that the alloy did not exhibit serrations in the plastic portions of stress–strain hysteresis loops in the temperature range examined. However, serrations were noticed during LCF deformation at lower strain rate ð3  104 s1 Þ at 773 K. In tensile deformation at a strain rate of 3  104 s1 , serrations were observed between 523 and 673 K. Other manifestations of DSA seen in tensile deformation were a hump in the yield stress and a plateau in the UTS [40]. In the present investigation, since the fatigue deformation is taking place at a higher strain rate of 3  103 s1 , the operative temperature range for DSA could get shifted to higher temperature values. It is inferred that the fatigue life decreases due to oxidation at strain rates less than 3  103 s1 at 823 and 873 K and due to combined effects of DSA and oxidation at 773 K (Fig. 8). Under conditions of low temperatures and high strain rates, oxidation effects were not significant. This is

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

166

illustrated in Fig. 9 where the fatigue crack propagation at a strain rate of 3  102 s1 at 823 K is depicted. Extensive branching of transgranular cracks is noticed on the tested samples associated with second phase boundaries especially at high strain amplitudes of testing. Further, microcracking perpendicular to the loading direction in the vicinity of the main crack was noticed at a strain rate of 3  102 s1 [10]. Hence the reduction of the fatigue life at high strain amplitudes and the higher strain rates could be rationalized based on the accelerated crack initiation and crack propagation associated with crack linkage observed in this alloy. The lower strain rates at elevated temperature provide adequate time for the environmental interaction to take place, which accelerates both the crack initiation and propagation phases (Fig. 8). The detrimental effect of oxidation was reflected in transgranular crack initiation and stage I and stage II crack propagation. In modified 9Cr–1Mo steel, the oxide layer formed on the surface of the

NUMBER OF CYCLES TO FAILURE

1200 773 K 823 K 873 K

3.3. Role of microstructure and phase transformation in LCF of stainless steel welds

1000

800

600

400

specimen is a weak barrier that can be easily overcome by slip. Ebi and McEvily [25] studied the influence of air environment on stage II crack growth in modified 9Cr–1Mo steel by comparing the air and vacuum test results. It was found that the surface of the specimen tested in vacuum was much more rumpled compared to that when tested in air. Also, several crack initiation sites were observed in the air-tested specimen whereas in vacuumtested specimens secondary cracks were found to be less. Challenger and Miller proposed a fall in threshold for crack propagation associated with the brittle oxides formed in a 2.25Cr–1Mo steel [41]. Ogata and Nitta [42] have used a similar concept, coupled with the wedge effect of exfoliated oxide, to explain the oxidation-assisted cracking in modified 9Cr–1Mo steel under creep–fatigue interaction conditions. The reduction in the LCF life under high temperature and low strain rate conditions could thus be ascribed to the enhanced environmental effects. Both crack initiation and propagation were seen to be transgranular under all the testing conditions investigated. This is to be expected since the 9Cr ferritic steels are generally known to resist grain boundary cavitation and associated intergranular failure even under creep–fatigue conditions.

10 -3

10 -2 -1

STRAIN RATE (s )

Fig. 8. Fatigue life of modified 9Cr–1Mo at various strain rates.

Fig. 9. Crack propagation at 823 K, 3  102 s1 .

LCF behaviour is affected by the initial microstructure and the microstructural changes that occur during deformation at high temperature. The dependence of the peak tensile stress on the number of cycles and on total strain amplitude for 316 weld metal, 316(N) weld metal at 773 K and 316L(N)/316 weld joint at 873 K is depicted in Figs. 10(a)–(c). Weld metal undergoes a relatively short initial hardening followed by a continuous softening regime without any apparent saturation period (Figs. 10(a) and (b)). The weld joints also displayed an initial hardening followed by a softening regime at all strain amplitudes, except at low amplitudes where a saturation stage was observed (Fig. 10(c)). In order to elucidate the operative deformation mechanisms detailed transmission electron microscopy of 316 weld metal was undertaken [4]. Untested samples revealed a very high dislocation density (Fig. 11(a)) and the dislocation structure was mainly tangles. The configuration of dislocations in the austenite matrix of the weld metal in the untested condition resembled that of a highly coldworked structure. This high dislocation density resulted from shrinkage stresses present during cooling of the weld metal during welding operations. No subgrains/cells were observed in the weld metal after fatigue testing. However, dislocation density after testing was much lower compared to as welded material, Fig. 11(b). This reduction in dislocation density could result from cyclic deformation which led to break down of dislocation tangles and subsequent annihilation of the dislocations.

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

+ 0.25% + 0.4% + 0.6% + 1.0% 300

200 316 WELD METAL TEMPERATURE: 773 K STRAIN RATE: 3x10-3s-1 100

(a)

101 102 103 NUMBER OF CYCLES

TENSILE STRESS AMPLITUDE (MPa)

TENSILE STRESS AMPLITUDE (MPa)

400 400

300

200

100

(b)

+ 0.4% + 0.6% + 1.0%

316(N) WELD METAL (N:0.07%) 773 K, 3 x10-3 s-1 101 102 103 NUMBER OF CYCLES

400 316L(N)/316 WELD JOINT Temperature : 873 K StrainRate :3 x 10-3 s-1 TENSILE STRESS AMPLITUDE / MPa

350

300

250

200

150 100

(c)

+ 1.0% + 0.6% + 0.4% + 0.25%

101 102 103 NUMBER OF CYCLES (N)

Fig. 10. Cyclic stress response of: (a) 316 weld metal, 773 K, (b) 316(N) weld metal, 773 K and (c) 316L(N)/316 weld joint, 873 K.

Fig. 11. (a) High dislocation density in austenite matrix before testing and (b) low dislocation density in austenite matrix after LCF testing.

167

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

The cyclic stress response of weld joints is marked by initial hardening (Fig. 10(c)), similar to the base metal although the degree of hardening is less. This behaviour could be justified since the major part of the gauge length is made up of the base metal, and consequently, the initial cyclic deformation of the weld joint would be similar to that of its major constituent. Furthermore, the occurrence of saturation at the lower strain amplitudes in the weld joint like the base metal suggested the predominance of deformation in the base metal or in the microstructurally altered region, i.e. the HAZ. A comparison of LCF lives of the base metal, weld metals and weld joints at 773 and 873 K revealed that at 773 K, the 316L(N) base metal showed a better fatigue resistance than 316 weld metal, Fig. 12(a). Further, 316(N) weld metal showed a lower fatigue endurance than 316 weld metal. However, at 873 K, 316 weld metal exhibited the highest fatigue resistance, Fig. 12(b). At both the temperatures 316L(N)/316(N) weld joints showed the least fatigue life. The variations in the fatigue life among the base metal, weld metal and weld joints with testing conditions can be correlated with the differences in the crack initiation and propagation behaviour. Metallographic observation of fatigue tested samples revealed that, in the base metal and 316 weld metal, crack initiation occurred in purely transgranular mode at 873 K, Fig. 13(a) [4]. However, in weld joints, crack initiation occurs intergranularly in the HAZ (Fig. 13(b)) where grain growth has occurred during welding. The decrease in life corresponded to a transition in crack initiation mode from purely transgranular mode to an intergranular one, in the weld joint. Fatigue crack propagation is transgranular in base metal and weld joints. In 316 weld metal, at 873 K, crack path deflection is noticed along austenite/ferrite interface (Fig. 13(d)), unlike at 773 K, Fig. 13(c). Further, in 316(N) weld metal, macrocrack propagation occurred without significant crack deflection (Fig. 13(e)), unlike 316 weld metal. The fine duplex austenite–ferrite microstructure in 316 weld metal with its many transformed phase boundaries at these testing conditions, offered a greater resistance to the extension of the fatigue cracks by causing deflection of the

PLASTIC STRAIN AMPLITUDE

-- -10 -2

316 WELD METAL 316(N) WELD METAL (N:0.07%) 316L(N) BASE METAL 316L(N)/316 WELD JOINT 316L(N)/316(N) WELD JOINT 316(N) WELD METAL (N:0.09%)

10 -3

10 -4

TEMPERATURE : 773 K -3 -1 STRAIN RATE : 3 x 10 s

10 2

(a)

10 3

10 4

crack paths compared to the base metal. d ferrite got transformed to brittle s phase during testing. This transformation is found to increase with increase in temperature (773–873 K) and decrease in strain rate, Tables 2 and 3. The extent of transformation was found to influence the fatigue life [4,7]. 316(N) weld metal with higher nitrogen content (0.09%) exhibited a better fatigue resistance than the one with lower nitrogen (0.07%), Figs. 12(a) and (b). This can be rationalized based on the influence of N on fatigue deformation of austenitic stainless steels. The beneficial effects of N on LCF life of austenitic stainless steels have been studied systematically by several investigators [13–19]. In these studies, nitrogen addition has been reported to be beneficial, and the LCF life has been improved at both ambient temperature and 873 K. However, this beneficial effect of nitrogen has been found to saturate at approximately 0.12 wt% of nitrogen. It must be pointed out that the addition of nitrogen promotes the occurrence of planar slip in austenitic stainless steels [5,18,43–45]. Generally, it is expected that addition of nitrogen in 316L steels reduce the stacking fault energy (SFE). However, based on detailed studies on the effect of nitrogen on SFE in these steels [18,45], it has been suggested that decrease in SFE with increase in N content is insignificant. From a detailed study of the effective stress and the internal stress components in cyclic deformation as a function of temperature and nitrogen content, it has been suggested that N enhances the pinning effect of dislocations thus favouring slip planarity [43]. In addition, short range order has been detected from TEM studies [46] on stress free thermally aged samples, in which evidences have been shown for the presence of ordered Cr2 N zones which could also contribute to slip planarity. Enhanced slip planarity increases the degree of reversibility of the fatigue deformation resulting in a reduced tendency for fatigue crack initiation and propagation, leading to an increased fatigue life. 3.4. Effect of hold time on fatigue life The introduction of strain hold at the peak strain in tension/compression in total strain controlled testing

-- -PLASTIC STRAIN AMPLITUDE

168

10 -2

10 -3

10 5

NUMBER OF REVERSALS TO FAILURE, 2Nf

316 WELD METAL 316(N) WELD METAL (N:0.07%) 316L(N) BASE METAL 316L(N)/316 WELD JOINT 316L(N)/316(N) WELD JOINT 316(N) WELD METAL (N:0.09%)

TEMPERATURE : 873 K -3 -1 STRAIN RATE : 3 x 10 s

10 2

(b)

10 3

10 4

NUMBER OF REVERSALS TO FAILURE, 2Nf

Fig. 12. Fatigue life of base metal and weld metals at: (a) 773 K and (b) 873 K.

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

169

Fig. 13. (a) Transgranular crack initiation and propagation in base metal, 873 K, (b) intergranular crack initiation in HAZ, weld joint, 873 K, (c) reduced crack deflection in 316 weld metal, 773 K, (d) crack deflection along the transformed d ferrite boundaries, 316 weld metal, 873 K and (e) reduced crack deflection in 316(N) weld metal, 873 K.

Table 2 Extent of transformation of d ferrite in 316 weld metal under different testing conditions Strain rates ðs1 Þ

Temperatures (K)

3  102 3  103 3  104

773

823

873

– – 24

21 30 64

32 40 76

Table 3 Extent of transformation of d ferrite in 316 and 316 (N) weld metals Temperatures (K)

773

873

Strain amplitudes (%)


0:25


0:4


0:6


1:0


0:25


0:4


0:6


1:0

316 weld metal 316(N) weld metal

2 –

– –

– –

– –

16 10

25 30

40 30

40 30

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

170

causes stress relaxation leading to creep–fatigue interaction. The creep–fatigue lives of the 316L(N) base metal, 316 weld metal and 316L(N)/316 weld joint as a function of the length of the hold time are shown in Fig. 14. It can be seen that, at 873 K, in both continuous cycling and hold time tests, 316 weld metal showed a higher fatigue endurance compared to the base metal. Further, the weld joint showed the lowest life. It is also evident that the hold time effect on fatigue life was dependent on the position as well as the duration of hold. Compared to continuous cycling conditions, imposition of hold at peak strain was found to decrease the fatigue life. Tensile hold was observed to be more damaging than the compression hold. Further, a significant reduction in fatigue life was observed by increasing the duration of tensile hold. The greater creep fatigue life of the weld metal compared to the base metal at 873 K can be correlated to the crack propagation differences between the base and weld metal, Fig. 15. Similar to continuous cycling tests under hold time, crack deflection at the transformed d phase boundaries is found to increase the crack propagation resistance causing an enhanced fatigue life. At a given strain amplitude, the relaxed stress during hold time at half-life for the 316L(N) base metal is provided in Table 4. In all the hold time tests, rapid stress relaxation occurs in the first few seconds of the strain hold, followed by a slower rate of stress relaxation during the rest of the hold period. During stress relaxation, conversion of elastic to plastic strain takes place and the strain rates are typically of the order of 104 –108 s1 during the slow relaxation period. The build up of tensile inelastic strain leads to the accumulation of grain boundary creep

Table 4 Effects of hold time on LCF properties at 873 K Hold (min)

D
t =2 (%)

Ds=2 (MPa)

sr (MPa)

N (cycles)

N=N f

0 1t 1c 10t 30t 90t 0 10t 30t 0 10t


0:6
0:6
0:6
0:6
0:6
0:6
1:0
1:0
1:0
0:4
0:4

328 307 323 291 288 274 370 342 322 280 247

– 51 45 57 74 94 – 68 96 – 40

580 ðN f Þ 475 510 409 330 235 130 ðN f Þ 140 110 680ðN f Þ 494

1 0.819 0.879 0.705 0.635 0.405 1 1.08 0.846 1 0.726

t: tensile hold; c: compression hold; D
t =2: total strain amplitude; Ds=2: half-life stress amplitude; N: number of cycles to failure (with hold); N f : number of cycles to failure (without hold) and sr : relaxed stress during hold.

Tension Hold ± 0.6%, Base Metal Comp. Hold, ± 0.6%, Base Metal

800

Fig. 15. Crack propagation in 316 weld metal, 1 min T hold, 873 K,
0:6%.

Tension Hold, ± 1.0%, Base Metal Tension Hold, ± 0.4%, Base Metal

NUMBER OF CYCLES TO FAILURE

Tension Hold, ± 0.6%, Weld Joint Tension Hold, ± 0.6% Weld Metal

600

400

200

0

316L(N) SS, 873 K, 3 x 10-3 s-1 0

20

40

60

80

100

HOLD TIME . min

Fig. 14. Fatigue life as a function of hold time.

120

damage in the form of cavities. With increase in the duration of the hold time a significant amount of stress relaxation takes place, leading to enhanced build up of intergranular creep damage. This conforms with the magnitude of sr developed during stress relaxation (Table 4) i.e. sr increases with increase in the length of the hold time and is greater in tension hold compared with the compression hold. It must be pointed out that the absolute magnitude of the sr alone cannot be associated with the damage that determines the creep–fatigue life. As a function of the strain amplitude, it is observed that sr is relatively large at high strain amplitudes. However, the degree of reduction in life during hold time tests, defined as N=N f (N, the fatigue life during hold-time tests and N f the corresponding fatigue life in continuous cycling) is found to be larger at lower strain amplitudes, compared to higher strain amplitudes of testing (Table 4). The strain rates during relaxation at higher strain amplitudes are generally higher than those observed at low strain amplitudes of testing. In general, the

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

relaxation strain rates of magnitude 4104 s1 observed at high strain amplitudes correspond to those which are expected to cause matrix deformation, while those observed at low strain amplitudes namely, p104 s1 , correspond to that of creep deformation. It has been suggested that relaxation strain rates p104 s1 generally contribute to grain boundary damage and cause a greater reduction in life [47,48]. The grain boundary damage developed during relaxation changes the modes of crack initiation and propagation. In 1 min compression hold, the crack initiation and propagation occur by transgranular mode similar to continuous cycling conditions, Fig. 16. However, under tension hold conditions crack initiation becomes intergranular and propagation is mixed mode (trans-+intergranular), Fig. 17. It is shown in Fig. 18(a) that the surface-connected slip bands are oxidized under 30 min tension hold conditions and the fracture surface is marked periodically by fatigue striations and intergranular cracks (Fig. 18(b)). Oxidation interaction is found to be more as the length of the hold time is increased to 90 min (Fig. 19). Both crack initiation and propagation seem to be strongly assisted by oxidation and fracture surface is completely covered by a thick oxide layer. The influence of the predominant fracture mode on fatigue can be assessed from a quantitative measurement of crack density i.e. cumulative length of the trans- and intergranular secondary cracks on the longitudinal section per unit area, as shown in Fig. 20(a) along with the frequency distribution of the crack length, Fig. 20(b). It is seen that intergranular crack density is the lowest in compression hold test, and crack propagation is mainly transgranular. Moreover, crack linkage, denoted by the number of large cracks ð4600 mmÞ, is also low in compression hold tests, Fig. 20(b). This behaviour can be attributed to the following reasons. During stress relaxation in compression hold, the growth of internal grain boundary cavities is very unlikely, since this needs both shear and normal tensile stresses across the grain bound-

Fig. 16. Transgranular crack propagation in 1 min compression hold, 316L(N) base metal.

171

Fig. 17. Mixed mode crack propagation in 1 min tension hold, 316L(N) base metal.

ary. Thus, for cycles containing compression hold periods, failure is dominated by transgranular crack initiation and growth mechanisms and fatigue lives are longer, compared with tensile hold periods. However, bulk creep damage is favoured under tension hold conditions. Crack density measurements on this steel also indicate that the transgranular crack density decreases with an increase in the length of the hold-time in tension and crack propagation is mainly intergranular. Moreover, it is noted that crack linkage becomes extensive as the length of the hold time increases and long cracks are observed in 10 and 90 min tension-hold tests, Fig. 20(b). Thus, the reduced fatigue life for longer hold-times could be ascribed to the occurrence of enhanced creep and oxidation damage at grain boundaries that facilitates accelerated intergranular crack initiation and propagation, and oxidation assisted transgranular fracture. 3.5. Thermo mechanical fatigue TMF is the primary life limiting factor for engineering components in many high-temperature applications [49,50]. TMF occurs, for example, in blades of gas turbines, in which depending on the location of the volume element considered and the cooling situation, various types of strain–temperature phasing can arise [49]. Further, the heating and cooling cycles during start-up and shut-down operations, such as those expected in high-temperature components of fast reactor, cause thermal stresses that often occur in combination with mechanical loads. The TMF cycles lead to material degradation mechanisms and failure modes typical of service conditions. The two extreme and basic experimental TMF cycle types that are most often used to assess life under TMF conditions are in-phase (IP) (peak tensile strain coincides with peak temperature) and out-of-phase (OP) (peak tensile strain coincides with the minimum temperature). The purpose of these studies is two-fold: first, to obtain engineering relationships and mathematical models for

ARTICLE IN PRESS 172

S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

Fig. 18. (a) Oxidized surface slip bands and (b) stage II propagation in 316 weld metal, 30 min tension hold, 873 K,
0:6%.

Fig. 19. Oxidation-induced intergranular cracking in 90 min tension dwell tests, 316L(N) base metal.

describing the macroscopic material behaviour, allowing design, evaluation and validation of engineering components and second, to gain deeper understanding of damage initiation and growth which are influenced by microstructure. Both IP and OP TMF tests were carried out on 316L(N) austenitic stainless steel on tubular samples using different mechanical strain amplitudes at a constant strain rate. The cyclic stress response of the materials (Fig. 21) displayed an initial hardening regime for the first 70–100 cycles followed by a well-defined saturation period that continues till the onset of crack initiation. This behaviour is typical of the alloy under strain-controlled cycling conditions. In contrast to strain-controlled isothermal LCF, the TMF cycling results in a mean stress, which is compressive under IP and tensile under OP cycling conditions, Fig. 22. The development of compressive mean stress in IP cycling is attributed to the decrease in the flow stress with increase in temperature during cycling from compressive peak to tensile peak strain and its increase with decrease in temperature during the reverse cycling. The converse effects operate in the OP cycling leading to the build-up of a tensile mean stress. In comparison with LCF at the peak temperature, the TMF cycling shows a higher stress response which can be attributed to the higher flow stress

of the material in the low-temperature regimes of the TMF cycle. Lives obtained under TMF and LCF cycling in the temperature range, 300–650 1C with increasing peak temperature on 316L(N) is provided in Table 5 [51]. The TMF life is seen to decrease with increasing peak temperature of thermal cycling. It was noted that TMF life in OP cycling is lower than that in the IP tests in the low-temperature regimes, whilst in the creep temperature domain, the IP tests yielded lower lives. The difference in endurance between the IP and OP cycling tests can be rationalized based on the cracking modes in the failed specimens. Metallographic examination showed that the mode of failure was transformed from the mainly transgranular in the lower temperature regimes for both IP and OP tests (Fig. 23) to a predominantly intergranular mode in the case of the IP tests (Fig. 24) where the peak temperature was beyond 600 1C. Influence of creep on the TMF damage was evident in the form of extensive intergranular cracking in the 400–650 1C IP test (Fig. 24). The isothermal tests conducted concurrently at the peak temperatures employed in the TMF cycles displayed lower lives compared to both IP and OP tests. The lower life of the alloy in the OP testing conditions in the lower temperature regimes could be attributed to the higher stress levels (Fig. 22) leading to an earlier crack initiation and a faster propagation. However, in the higher temperature ranges, IP cycling proved to be more deleterious as a consequence of the creep-dominated intergranular cracking that was seen to be aided by oxidation, Fig. 24(a). Influence of oxidation was more prominent in the IP tests as a consequence of the cracks remaining open at the high-temperature end of the TMF cycle, thereby allowing easy environmental access to the crack tips. Evidence of creep dominated intergranular cracking was clearly seen on the fracture surface of specimen tested under 400–650 1C IP cycling, whereas creep effects were absent in OP cycling conditions. Fractograph presented in Fig. 24(b) illustrates the creepinduced intergranular cracking interspersed with clear striations under the above testing condition. The isothermal tests conducted concurrently at the peak temperatures employed in TMF cycles displayed lower

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175 400

16

1 min. Tension

1 min. Tension

90 min. Tension

300

1 min. Compression

250 200 150 100

90 min. Tension

12

10 min. Tension 1 min. Compression

10 8 6 4

50 µm/mm2

14

10 min. Tension

NUMBER OF CRACKS

CRACK DENSITY

350

173

2 0

0

1 Transgranular

(a)

Intergranular

1-2

2-3

3-4

4-5

5-6

6-7

7-8

>8

CRACK LENGTH (× 100 µm)

(b)

Fig. 20. (a) Crack density as a function of hold time, (b) crack length distribution as a function of hold time 873 K.

300

300-550 C 350-600 C LCF, 550 C LCF, 600 C 400-650 LCF, 650 C

200

Stress, MPa

100 0 0

500

-100

1000

1500

2000

2500

3000

Number of cycles

-200 -300 -400

Fig. 21. Cyclic stress response obtained under TMF–IP and LCF cycling, D
mech ¼
0:4%.

Table 5 TMF and LCF life of 316L(N) stainless steel at various temperatures ðD
mech ¼
0:4%Þ

400 300

Temperature range, DT

200

Stress, MPa

300–550 350–600 400–650

In-Phase

100

Out-of-Phase

TMF life

Isothermal LCF life at maximum temperature ðT max Þ

In-phase

Out-of-phase

2093 1210 585

1312 1000 853

693 835 672

0 0 -100

500

1000

Number of cycles

-200 -300 -400

Fig. 22. Comparative stress response: IP and OP cycling at D
mech ¼
0:4%, 350–600 1C.

lives compared to both IP and OP tests and hence for austenitic stainless steels LCF results can be safely used for design. It must be pointed out that isothermal tests do not always provide useful and complete information necessary to enable validation of design and life prediction of components subjected to TMF loading. This is particularly so in materials having different deformation mechanisms operating at different temperature, materials exhibiting ductile-to-brittle transitions and materials with thermal

ARTICLE IN PRESS 174

S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175

Fig. 23. Crack propagation under (a) 300–550 1C IP cycling, (b) 350–600 1C OP cycling.

Fig. 24. (a) Oxidation-assisted mixed mode cracking under 400–650 1C IP cycling, (b) SEM fractograph showing creep-dominated IG cracking interspersed with clearly defined striations.

barrier coatings whose integrity need to be assessed under TMF loading conditions. In these materials due to reduced life observed in TMF, it is important to model the process accurately from isothermal fatigue tests. However, depending on the actual loading conditions, damage evolves differently in TMF and LCF and life prediction may be non-conservative, if relevant damage mechanisms are not taken into account. Studies that assess the predictive capabilities of TMF models are rare [49]. Life prediction under TMF from isothermal LCF tests is a challenging task. 4. Conclusions DSA was found to influence cyclic hardening in 316L(N) SS and modified 9Cr–1Mo steel in the sub-creep temperature range. The important manifestations of DSA include serrations on the stress–strain hysteresis loops, a negative strain rate–stress response and inverse temperature dependence of saturation stress. Further, enhanced slip planarity associated with DSA caused an increase in the intergranular crack density and a reduced fatigue life. Cyclic softening observed in weld metals was attributed to the annihilation of dislocations. 316L(N) base metal exhibited better fatigue resistance than 316 weld metal at 773 K. However, at 823 and 873 K, at specific strain rates, 316 weld metal showed the highest fatigue life. 316(N) weld metal possessed lower fatigue life compared to 316 weld metal. Further, weld joints showed the lowest life under all testing conditions. Better fatigue life of 316 weld metal was attributed to a high degree of

crack deflection associated with the transformed d ferrite boundaries. The poor fatigue resistance of the weld joints was attributed to intergranular crack initiation and poor crack propagation resistance of the coarse-grained HAZ. At lower strain rates and long hold times at high temperatures, oxidation was found to influence fatigue life in 316L(N) SS and modified 9Cr–1Mo steel. The greater creep–fatigue life of the 316 weld metal compared to the base metal at 873 K under hold time testing conditions can be correlated to the crack propagation differences between the base and weld metals. The TMF life of 316L(N) stainless steel was seen to decrease with increasing peak temperatures of thermal cycling. The isothermal tests conducted at the peak temperatures of the thermal cycles displayed lower lives compared to both the IP and the OP cycling. Further, lives under OP cycling were found to be lower than those under IP conditions in the low-temperature regimes, while the converse holds good when the upper temperature encompassed the creep–dominant regime. Influence of creep on the TMF damage was evident in the form of extensive intergranular cracking in the 400–650 1C IP test. References [1] Hales R. ASME-PVP, vol. 70. 1982. p. 1. [2] Degallaix S, Degallaix G, Foct J. Low Cycle Fatigue. ASTM STP 942; 1988. p. 798. [3] Srinivasan VS, Sandhya R, Bhanu Sankara Rao K, Mannan SL, Raghavan KS. International Journal of Fatigue 1991;13(6):471. [4] Valsan M, Sundararaman D, Bhanu Sankara Rao K, Mannan SL. Metallurgy and Material Transactions 1995;26A:1207.

ARTICLE IN PRESS S.L. Mannan, M. Valsan / International Journal of Mechanical Sciences 48 (2006) 160–175 [5] Srinivasan VS, Valsan M, Bhanu Sankara Rao K, Mannan SL, Sastry DH. Transactions of the Indian Institute of Metals 1996; 49:489. [6] Srinivasan VS, Valsan M, Bhanu Sankara Rao K, Mannan SL, Sastry DH. International conference on pipes and pressure vessels, Singapore, 1996. p. 255. [7] Nagesha A, Valsan M, Bhanu Sankara Rao K, Mannan SL. In: XueRen Wu, Wang ZG, editors. Proceedings of the seventh international fatigue conference FATIGUE ’99, vol. 2. Beijing, China; 1999. p. 1303. [8] Srinivasan VS, Valsan M, Sandhya R, Bhanu Sankara Rao K, MannanL SL, Sastry DH. International Journal of Fatigue 1999;21:11. [9] Nagesha A, Valsan M, Kannan R, Bhanu Sankara Rao K, Mannan SL. International Journal of Fatigue 2002;24(12):1285. [10] Mannan SL, Bhanu Sankara Rao K, Valsan M, Nagesha A. Proceedings of the fourth conference on creep, fatigue and creep–fatigue interaction, October 8–10, Kalpakkam; 2003. p. I-157. [11] NRIM fatigue data sheet, No. 78, NIMS, Japan 1993. [12] Srinivasan VS, Valsan M, Bhanu Sankara Rao K, Mannan SL, Raj B. International Journal of Fatigue 2003;25:1327. [13] Taillard R, Degallaix S, Foct J. In: Rie K-T, editor. Low cycle fatigue and elasto-plastic behaviour of materials. Munich: Elsevier; 1987. p. 83. [14] Sandstrom R, Engstrom J, Nilsson JO, Nordgren A. High Temperature Technology 1989;7:2. [15] Nilsson JO. Fatigue and Fracture Engineering Material Structure 1894;7(1):55. [16] Vogt B, Degallaix S, Foct J. International Journal of Fatigue 1984; 6(4):211. [17] Nilsson JO. Scripta Metallurgica 1983;17:593. [18] Taillard R, Foct J. In: Foct J, Hendry A, editors. Proceedings of the international conference on high nitrogen steels 88. The Institute of Metals and the Socitie francaise de Metallurgie, Lillie, France; 1988. p. 387. [19] Nilsson JO, Thorvaldsson T. Scandinavian Journal of Metallurgy 1985;15:83. [20] ASME Boiler and Pressure Vessel Code ASME Section III, Division I, sub section NH, New York; 1998. [21] RCC-MR Appendix Z—A9 1J.4. [22] Brinkman CR, Gieseke BG, Maziasz PJ. In: Liaw PK, Viswanathan R, Murty KL, Simonen EP, Frear D, editors. Microstructure and mechanical properties of ageing material. The Minerals, Metals and Materials Society; 1993. p. 107. [23] Sikka VK. In: Davis JW, Michel DJ, editors. Proceedings of topical conference on ferritic alloys for use in nuclear energy technologies. Snowbird UT, June 19–23; 1983. p. 17. [24] Sanderson SJ. In: Khare AK, editor. Proceedings of ASM international conference on production, fabrication, properties and application of ferritic steels for high temperature applications. Warren, PA, October 6–8, 1981. Metals Park, OH: ASM; 1983. p. 85. [25] Ebi G, McEvily AJ. Fatigue Fracture Engineering Material Structure 1984;17:299. [26] Sikka VK, Ward CT, Thomas KC. Ferritic steels for high temperature applications. In: Khare AK, editor. Proceedings of ASM international conference on production, fabrication, properties and application of ferritic steels for high temperature applications, Warren, PA, October 6–8, 1981. Metals Park, OH: ASM; 1983. p. 65.

175

[27] Gieseke BG, Brinkman CR, Maziasz PJ. In: Liaw PK, Viswanathan R, Murty KL, Simonen EP, Frear D, editors. Microstructure and mechanical properties of ageing material. The Minerals, Metals and Materials Society; 1993. p. 197. [28] Swindman RW. Low cycle fatigue. In: Solomon SD, Halford GR, Kaisand LR, Leis BN, editors. ASTM STP 942. Philadelphia, PA: ASTM; 1998. p. 107. [29] Asada Y, Dozaki K, Ueta M, Ichimiya M, Mori K, Taguchi K, et al. Nuclear Engineering Design 1993;139:269. [30] Jones WB. Ferritic steels for high temperature applications, In: Khare AK, editor. Proceedings of ASM international conference on production, fabrication, properties and application of ferritic steels for high temperature applications, Warren, PA, October 6–8, 1981. Metals Park, OH: ASM; 1983. p. 221. [31] Eifler D, Rottger D. In: Xue-Ren Wu, Zhong-Guang Wang, editors. Proceedings of the seventh international fatigue congress, Fatigue ’99, June 1999, Beijing, vol. 4; 1999. p. 2145. [32] Nishino S, Shiozawa K, Takahashi K, Seo S, Yamamoto Y. In: XueRen Wu, Zhong-Guang Wang, editors. Proceedings of the seventh international fatigue congress, Fatigue ’99, June 1999, Beijing, 1999. p. 2177. [33] Kim S, Weertman JR. Metallurgy Transactions A 1988;19:999. [34] Okamura H, Ohtani R, Saito K, Kimura K, Ishii R, Fujiyama K, et al. Nuclear Engineering Design 1999;193:243. [35] Sugiura T, Ishikawa A, Nakamura T, Asada Y. Nuclear Engineering Design 1994;153:87. [36] Aoto K, Komine R, Ueno F, Kawasaki H, Wada Y. Nuclear Engineering Design 1994;153:97. [37] Zhu L, Zhao Q, Gu H, Lu Y. Journal of Material Science and Technology 1998;14:226. [38] Mannan SL. Bulletin of Material Science 1993;16:561. [39] Valsan M, Sastry DH, Bhanu Sankara Rao K, Mannan SL. Metallurgy Transactions 1994;25A:159. [40] Samuel KG, Bharathi G, Bhanu Sankara Rao K, Mannan SL. Internal Report, PFBR, MPS/Modified 9Cr–1Mo, 1998. [41] Challenger KD, Miller AK. Journal of Engineering Material Technology ASME 1981;103:7. [42] Ogata T, Nitta A. Proceedings of the 30th symposium on structural materials at high temperature; 1992. p. 149. [43] Vogt JB, Magnin T, Foct J. Fatigue and Fracture Engineering Material Structure 1993;16(5):555. [44] Ermi AM, Monteff J. Metallurgy Transactions 1982;13A:1577. [45] Vogt JB, Magnin T, Foct J. Proceedings of the fourth international conference on fatigue and fatigue thresholds, Fatigue ’90; vol. 1. 1990. p. 87. [46] Shankar P, Sundararaman D, Ranganathan S. Scripta Metallurgica 1994;31(5):589. [47] Bhanu Sankara Rao K, Meurer HP, Schuster H. Material Science Engineering 1988;104A:37. [48] Rie KT, Schmidt RM, Ilschner B, Nam SW. ASTM STP 942, Philadelphia, 1988. p. 313. [49] Christ H-J, Maier HJ, Teteruk R. Fourth conference on creep, fatigue and creep–fatigue interaction, October 8–10, Kalpakkam; 2003. p. I–11. [50] Schallow P, Christ H-J. Fourth conference on creep, fatigue and creep–fatigue interaction, October 8–10, Kalpakkam; 2003. p. C-51. [51] Nagesha A, Valsan M, Bhanu Sankara Rao K, Kannan R, Mannan SL. Fourth conference on creep, fatigue and creep–fatigue interaction, October 8–10, Kalpakkam; 2003. p. C-253.