National Research Institute for Metals has proceeded the fatigue data sheet project, in which fatigue properties have been examined systematically for engineering materials manufactured in Japan. The results obtained have been published as NRIM Fatigure Data Sheets. In this paper, the effects of temperature and stress concentration factor on high cycle fatigue properties were summarized for nine kinds of engineering steels, fatigue data of which were published in NRIM Fatigue Data Sheets. The conclusions obtained were as follows: (1) The ratio of the fatigue strength of the smooth specimens at 108 cycles to the tensile strength at the same temperature varied a little ranging from 0.4 to 0.6 depending on temperature and material conditions. The mean value was about 0.5. (2) The notch effect on fatigue strength varied depending on temperature and material conditions. The reason is considered as that the fatigue strength of the notched specimens for failure has a tendency to become larger than that for crack initiation at high temperatures. (3) The fatigue strength of the notched specimens at 108 cycles for crack inititation corresponded well to the tensile strength independent of temperature and material conditions. (4) The notch sensitivity index η had a tendency to become small with increasing stress concentration factor or with decreasing tensile strength.
In order to clarify the frequency effect on high cycle fatigue properties at intermediate temperatures, rotating bending fatigue tests were carried out on a carbon steel S45C at 200, 300, 400 and 500°C with frequencies of 10 and 1Hz. The conclusions obtained were as follows: (1) The fatigue strength at 300°C was larger than that at 200°C, and a negative frequency effect on fatigue strength was observed at this temperature range. At above 400°C, the fatigue strength decreased with increasing test temperature and a positive frequency effect was observed. (2) The frequency dependence of fatigue strength corresponded with that of deformation stress. The negative frequency dependence of fatigue strength observed at 300°C was considered as the result of decreasing cyclic plastic strain at the surface of specimen with decreasing frequency by hardening on account of dynamic strain aging. (3) The positive frequency dependence of fatigue strength observed at 500°C was considered as the result of increasing cyclic plastic strain at the surface of specimen with decreasing frequency by creep deformation and by softening on account of tempering effect.
The stress-strain-temperature relation in TiNi alloys was investigated from the phenomenological point of view. A set of constitutive equations consisting of the mechanical constitutive equation and of the transformation kinetics were presented to describe the transformation pseudoelasticity, the partial pseudoelasticity and the shape memory effect. The material constants in the theory were determined from the tension tests on TiNi wire specimens at various temperatures. The theory was successfully applied to the calculation of the recovery force induced in the heating process under the condition of constraint strain.
Biaxial loading tests were carried out by subjecting a thin tube of SUS316 steel to cyclic torsional straining at 923K under a constant axial load. The aim of the tests was to observe and characterize the ratchetting deformation at elevated temperature under multiaxial loading. The results obtained were as follows: (1) The ratchetting strain εa was proportional to the constant axial stress σ and the power of the number of cycles N, in such a way as εa=C·σ·Nα. (2) The effect of the cyclic torsional strain rate γ on the ratchetting strain εa was approximately represented by εa=C'·γβ. (3) The growth rate of the ratchetting strain slowed down with decreasing strain amplitude, and the shakedown behavior was observed at Δγ/√3=0.5%. (4) The classical plastic analysis of the ratchetting deformation predicted an overestimation of accumulation of the ratchetting strain. A simplified ratchetting analysis method based on the effective primary stress concept gave good correlation with the experimental results.
Description of anisotropic damage state in terms of mechanical variables and its application to anisotropic creep damage theory were discussed. The notion of the classical Kachanov-Rabotnov theory of creep damage was first extended to the general three-dimensional state of damage by introducing fictitious undamaged state and the fictitious deformation on gradient. Then, these results were applied to formulate the evolution and the constitutive equations of creep crack growth analysis of copper plates under proportional and non-proportional biaxial loadings.
Crack propagation under sustaining or increasing load conditions in monotonic creep and time-dependent fatigue has been described successfully by the creep J-integral. This paper extends the creep J-integral concept to the crack propagation in decreasing load conditions. Because the J-integral can be originally defined for a linear or nonlinear elastic body even in an unloading condition although it can not for a plastic body, and because the creep J-integral in a creep body is analogous to the J-integral in an elastic one, the creep J-integral can be expected to be valid in the decreasing load (stress relaxation) condition when a unique stress-strain rate relation exists. Crack propagation tests of 0.16% carbon steel were carried out in creep and time-dependent fatigue conditions with the stress relaxation. As the result, the crack propagation rate had a good correlation with the creep J-integral for the stress relaxation period. In time-dependent fatigue, moreover, the crack propagation rate was correlated well with the creep J-integral (or creep J-integral range) independently of the stress waveform. These show the applicability of the creep J-integral concept for crack propagation during the stress relaxation.
The formation mechanism of surface cracks leading to creep fracture was investigated through microscopic observations and measurements of grain boundary sliding during creep, in a 321 stainless steel. Creep tests were carried out in argon containing 4% hydrogen at 923K under the stresses of 196, 176 and 157MPa. The grain boundary sliding during the creep tests were measured by the displacements of scratched lines and interference fringes at the grain boundaries on the crept specimen's surface. The grain boundary displacements could be detected clearly after the primary creep stage, and increased in proportion to the creep strain during the secondary stage and the initial part of the tertiary stage. The surface cracks were formed when the grain boundary displacement reached a critical value of about 0.3μm. This suggests that the surface cracking depends mainly on the grain boundary sliding. The microscopic observation indicates that the surface cracking process is as follows. (1) Small cavities form at the grain boundaries on surface. (2) The cavities grow along the grain boundaries and become crack-like due to their coalescence. (3) The crack-like cavities grow into cracks of grain size and wedge-like shape. This process is caused mainly by the grain boundary slinding.
This paper describes the change in mechanical properties of steels after long-term service in power boilers and the evaluation of creep damage for boiler tubing. Interrupted creep tests were carried out for SUS321HTB, STBA26, STBA24 tubes after the service exposure and followed by metallographic studies. The following conclusions were obtained. (1) The change in mechanical properties after the service exposure is related with the change in inherent structure of the steels. In the case of boiler tubing, these changes became very slight at the longer period. (2) For SUS321HTB, creep voids were observed on the grain boundaries at the end of steady state creep. The density change of the crept specimens had a good relation to the creep strain. (3) It was confirmed that the ferritic steels such as STBA24 and STBA26 are very ductile and hard to form voids and cracks even at the accerelated creep. (4) Based on these results, a life evaluation system for boiler tubing was proposed.
Safety margins in boiler tube design are severely lowered by either local heating or corrosion during operation in some cases. For controlling the safety margins, it is popular to conduct metallog-raphical inspection via replica, tube temperature measurement in doghouse and tube thickness measurement. Recently, electrochemical test and density measurement were proposed, which had been rarely applied previously due to the uncertainty in the correlationship between material property change and life consumption. We made efforts to improve the conventional boiler tube life prediction method into a more practical one. The results obtained are summarized as follows: (1) The extrapolation program (error: 0.8-1.2) developed in this study can estimate the creep-rupture stength of steels for decade-long use on the basis of test data up to 3000 hours. (2) By using the corrosion rate, the degree of local overheating and material deterioration data, the new program can estimate the residual life of boiler tubes in service in error of a factor of three. (3) The life prediction method was applied to four superheaters and one reheater. In two cases, the residual lives were found to be very short.
Since the reformer-tubes used in fuel-cell systems are subjected to frequent load- and temperature-cycling in the creep range, it is necessary to take account of creep-fatigue interaction for material selection and structural design. In order to find the most suitable materials and life estimation method for this application, creep rupture and low-cycle-fatigue properties at 800 and 1000°C were investigated on various centrifugally-cast materials. The following conclusions were obtained; (1) A Nb- and Ti-containing 25Cr-35Ni steel (HP-BST-M) was superior in creep rupture strength and creep-fatigue properties to conventional HK40 and HP-Nb at both 800 and 1000°C. (2) A low carbon HP-Nb showed equally as good creep-fatigue properties as HP-BST-M had at both temperatures. This material had almost the same creep rupture strength as HK40 had at 800°C, but had higher strength than HK40 at 1000°C. (3) The superior creep-fatigue properties of HP-BST-M and low carbon-HP-Nb were attributed to the excellent creep ductility. (4) The linear cumulative damage rule (φc+φf=1) gave the extremely conservative life estimation at 800°C, but gave the unconservative one at 1000°C. On the other hand, the strain-range-partitioning method gave an accurate prediction irrespective of temperature by selecting appropriate partitioning.
An investigation of steam turbine components retired after long-term service at elevated temperatures has clarified that one of the most typical characteristics of material degradation in Cr-Mo-V steel is softening, and that the mechanical properties necessary for life prediction such as creep and low cycle fatigue are affected by material degradation (softening). As a typical example, this paper shows the quantitative relationships between the material hardness and these degraded mechanical properties for Cr-Mo-V rotor forgings. Creep and fatigue life assessment for steam turbine rotors could be more accurately made through the nondestructive measurement of hardness and the evaluation of the degraded creep and fatigue properties. A life diagnosis system was developed based on the present procedure with a 32 bit super mini-computer and man-machine interfaces. This system helps us perform easier and more sophisticated life diagnosis and maintenance of steam turbine main components used at high temperatures such as rotors, casings, valves, blades and bolts.
In order to investigate the effect of previous creep damage on the fatigue crack propagation behavior at elevated temperature, stress-controlled fatigue crack propagation tests were carried out at 650°C on SUS304 steel. The results obtained were summarized as follows: (1) For the fatigue crack propagation test, the data of crack propagation rate could be arranged by ΔJ. As a result, the crack propagation rate of the previously creep-damaged material was half of that of the virgin material when the degree of previous creep damage and the creep stress level were larger than some critical values. This may be caused by the change in microstructure of the material, that is, the introduction of the subgrain boundaries due to the strong cell formation during the previous creep loading. (2) For the creep-fatigue test, the data of crack propagation rate could be arranged by ΔJc. As a result, the crack propagation rate of the previously creep-damaged material was almost same as that of the virgin material. It seems that the crack propagation rate was not affected by the change in microstructure inside the grain due to the previous creep loading because the creep-fatigue crack propagated along the grain boundaries.
Fatigue tests in sodium fluid at elevated temperatures were conducted to investigate the effect of sodium on low-cycle fatigue properties of SUS304 stainless steel, which is the main structural material in FBRs. The tests were conducted at 500°C, 550°C and 600°C under the following condition; the sodium flow rate of 1m/sec, the cold trap temperature of 120°C, the strain rate of 1×10-3/sec, and the total strain range from 0.3 to 1.9% with trianglar wave form. It was found that the fatigue lives in sodium were equal to or greater than those in air. Secondary cyclic hardening, which is considered to be caused by precipitation, was observed at 500°C and 550°C in the low strain range. This phenomenon was evident in sodium at 500°C and was restrained by pre-precipitation which occured during aging. Crack propagation was transgranular and striations were observed on the fracture surfaces. From the qualitative crack propagation analysis based on the measured striation spacing, it was shown that the propagation rate in sodium was almost the same as that in air. It was presumed that the difference in fatigue life was mainly caused by the difference in crack initiation time. Surface cracks found in sodium were fewer than those found in air. This observation shows that cracks are more difficult to initiate in sodium than in air. The reduction in crack formation was considered to be caused by the suppression of oxidation by the reducing atmosphere. It was presumed that the oxidation suppression caused the increase in crack initiation time.
In order to investigate the properties of biaxial low-cycle fatigue in SUS304 stainless steel at elevated temperature, strain controlled, tension-compression-torsion fatigue tests were carried out under in-phase and out-of-phase conditions between axial and torsional strain cyclings. The fracture mode under the in-phase cycling was found to be classified into two types; i.e. Mode I and Mode II. It was also found that the fracture mode changed from Mode I to Mode II with an increase in the strain ratio, Δγ/Δε, and the transition occured at about 1.7 of Δγ/Δε. On the other hand, under the out-of-phase cycling, the mixed failure of Mode I and Mode II was found from SEM fractographic observation. The in-phase fatigue life with the Mode I type failure was correlated well with the tensile strain energy and that with Mode II type failure the torsional strain energy. By using these correlations, the out-of-phase fatigue life could be predicted by the linear damage rule.
This paper describes the low cycle fatigue behavior of 1Cr-1Mo-1/4V and SUS304 cruciform specimens in wide range biaxial stress conditions of -1.0≤φ≤1.0 at high temperatures, where φ denotes the ratio of the minimum principal strain to the maximum principal strain existing on the specimen surface. The detailed description of the newly developed biaxial fatigue testing machine was made, and the biaxial test results were discussed in connection with the biaxial low cycle fatigue parameter. The COD and ther Γ*-plane parameters were most appropriate to express the biaxial fatigue behavior. Also the X-ray diffraction was made in order to examine the damage in biaxial stress states.