In a previous work, the present authors proposed a new extension of the constitutive model of cyclic plasticity developed by Ohno and Wang, and they verified the extension on the basis of uniaxial ratchetting experiments of 316FR steel at room temperature. The extension is featured by the capability of representing steady-state ratchetting. In this work, the extension is examined further for the validity under nonproportional cyclic loading by simulating the viscoplastic behavior of IN738LC subjected to multiaxial ratchetting, multiaxial cyclic stress relaxation, butterfly-type cyclic straining, and circular cyclic straining with strain hold at 850°C. The original Ohno and Wang model and the Armstrong and Frederick model are also examined. It is shown that the extension enables us to simulate accurately the multiaxial viscoplastic behavior mentioned above, and that if neither ratchetting nor mean stress relaxation under cyclic loading is significant the three models provide nearly the same predictions as one another.
An welded joint of 304 stainless steel was fabricated by SAW with use of 308 filler metal to investigate the influence of welding pass sequence and sampling location of the specimen in the welded joint on the creep properties of weld metal zone and heat affected zone in the joint. Creep tests were conducted at 823K and 235MPa. In the HAZ of the welded joint, there existed a hardened region of which specimen showed a little smaller creep rate and longer life than base metal. The 308 weld metal had larger creep rate than the 304 base metal and the life of the former was only 10% of that of the latter (under-matching), however, the difference of the lives becomes smaller for long times. 0.2% proof stress and creep properties of the weld metal zone depended on the sampling location in the welded joint due to the effect of the weld pass sequence and the restraint used during welding. The most important factor that affects the creep properties of weld metal zone would be the difference of micro structure among the beads which was generated from the pass sequence on multi-layer welding. From the results mentioned above, it is necessary, on evaluating the strength of the joint and predicting its life by simulation, to take the heterogeneity of the mechanical properties within the welded joint into account.
Using hot-rolled 316FR stainless plate (50mm thick) and 16Cr-8Ni-2Mo filler wire, a narrow-gap welded joint was prepared by GTAW (gas tungsten arc welding) process. In addition to conventional round bar specimens of base metals and weld metal, full-thickness joint specimens were prepared for creep test. Creep tests were conducted at 550°C in order to examine creep deformation and rupture behavior in the weld metal of the welded joint. Creep strain distribution on the surface of the joint specimen was measured by moiré interferometry. In the welded joint, creep strength of the weld metal zone apart from the surface was larger than that in the vicinity of the surface due to repeating heat cycles during welding. Creep strain and creep rate within the HAZ adjacent to the weld metal zone were smaller than those within the base metal zone. Creep rate of the weld metal zone in the welded joint was smaller than that of the weld metal specimen due to the restraint of the hardened HAZ adjacent to the zone. The full-thickness welded joint specimens showed longer lives than weld metal specimens, though the lives of the latter was shorter than those of the base metal (undermatching). In the full-thickness welded joint specimen, crack started from the last pass layer of the weld metal zone and fracture occurred at the zone. From the results mentioned above, in order to evaluate the creep properties of the welded joint correctly, it is necessary to conduct the creep test using the fullthickness welded joint specimen which includes the weakest zones of the weld metal, the front and back sides of the plate.
The relation between creep rupture properties and microstructure of weld metal on 2.25Cr-1Mo steel was investigated. Small round bar specimens were prepared from the weldment performed by multi-layer welding. The creep rupture tests of the weld metal specimen were conducted at 550°C and the stress ranged from 69 to 159MPa under the constant load condition. The fracture mode of weld metal specimen was examined and related to the microstructure. The major conclusions are: (1) Three kinds of microstructure were observed in the weld metal performed by the multi-layer welding: columnar structure due to rapid cooling, microstructure due to reheating by subsequent passing, and a similar bainitic microstructure to that of the base metal arising from cyclic heating. The difference was not observed in the creep rupture properties of the weld metal specimens sampled from different locations in the plate. (2) Creep rupture lives become larger in the order of the base metal, the weld metal and the welded joint at the stresses higher than 98MPa, however the difference becomes smaller at the lower stresses and the long times. (3) The decrease ratio of the hardness of the grip head in the specimen ruptured after 77845h to that of the aswelded specimen reached only 10%, while the ratio of the hardness of the gauge portion in the same specimen was 30%. The decrease strongly depended on the applied stress and the cause of the decrease would be ascribed to agglomeration and coarsening of carbides or recovery of the dislocation substructure. (4) As the applied stress increased, fracture mode of the weld metal specimens varied from ductile fracture, via the fracture at columnar interface, to the mixed fracture at the columnar interface and at prior austenitic grain boundaries.
The creep-fatigue life prediction model previously proposed for evaluating Mod.9Cr-1Mo steel under variable strainings was applied to 316LC steel. In the proposed model the small crack growth curves in a smooth specimen are predicted in accordance with the applied straining conditions by using the material parameters and they are used as the damage accumulation curves. The feature of this model is its potential to evaluate the material damage and the residual life under arbitrary variable creep-fatigue strainings. For examining whether or not it is possible to determine the material parameters for 316LC steel based on the same creep-fatigue life prediction model as for Mod.9Cr-1Mo steel, creep-fatigue tests, crack growth tests, and high-low (HL) and low-high (LH) type two-step variable straining tests were conducted with PP (fast-fast) and CP (slow-fast) type strain waveforms at 700°C and 800°C. It was suggested that the crack initiation life cannot be negligible not only in PP type low strain ranges but also in CP type low strain ranges whereas it is negligible in CP type low strain ranges for Mod.9Cr-1Mo steel, and the model was partly modified to describe the initial crack growth behavior of 316LC steel. The remaining life in HL and LH tests was well predicted by the modified proposed model.
Under the circumstance that aging of boiler components proceeds and some “Type IV” cracking incidents were recently reported in boiler weldment parts, development of an accurate remaining life evaluation method for the weldment parts is an important subject for the utilities. In this study, in order to clarify a cause of the “Type IV” cracking under creep-fatigue loading condition and to develop a life evaluation method for boiler weldment parts, creep-fatigue tests on the heat affected zone (HAZ) simulated materials, base metal, weld metal and weld joint of 2.25Cr-1Mo steel, and elastic-plastic and creep analysis for the weld joint was conducted. As a result, initiation of many cavities and “Type IV” cracking were observed in a fine grain region of the weld joint specimen failured under the creep-fatigue loading. It was found from the comparison between experimental evidences and the analytical results that “Type IV” cracking was caused by two major reasons. One of them is accumulation of creep strain during strain hold in the fine grain region is larger than that in other region suggesting progress of creep damage in the fine grain region prior to other regions. The other one is existence of multiaxial tensile field within the fine grain region caused reduction of failure ductility. Crack initiation portion and failure life under the creep-fatigue test could be well predicted by the non-linear damage accumulation model based on finite element analysis using conventional elastic-plastic and creep theory.
The idea and procedures are presented for evaluating the remaining life based on a new creep-fatigue damage rule. The proposed method needs only information about crack length at a given number of cycles or crack length change during a given cycles interval and Fij, the ratio of the partitioned inelastic strain range to the total inelastic strain range. It is shown that the material damage and the applied inelastic strain range as well as the remaining life of IJ tested Mod.9Cr-1Mo steel can be evaluated when the difference between the measured crack length and the hypothetical initial crack length or the measured crack length change is greater than 30μm, and that the respective accuracy in predicting each of the remaining life, the applied inelastic strain range and the material damage is within about factor of 3-4, 2.5-2.8 and 1.5-3.
The Incoloy 908 jacket material exhibited intergranular cracking during superconductor reaction heat treatment at 650°C. The mechanism of this intergranular cracking was attributed to stress-accelerated grain boundary oxidation (SAGBO). Incoloy 908 is the jacket material for the cable-in-conduit superconductor for the International Thermonuclear Experimental Reactor (ITER). Fracture surfaces mapped by wave dispersive X-ray (WDX) analyses indicated an accumulation of oxygen at the crack initiation region, but did not show any presence of oxygen around the crack tip. A computational simulation of the conductor manufacturing process indicated the presence of residual surface stresses in the jacket material. Residual surface tensile stresses are created on the jacket outer surface, as a result of plastic deformation during conductor fabrication (known as the jacket compaction process). It has been concluded that the successful heat treatment of the jacket material can be accomplished by: (1) Careful shotpeening on the outer surface of the jacket material to create compressive surface stresses and (2) Evacuation of oxygen by removing oxygen sources such as the oil contaminants and water vapor from the conductor at temperatures below 550°C.
Diffusion bonding of an oxide dispersion strengthened (ODS) Ni-base alloy, MA 758, was studied by combining several bonding parameters: bonding temperature, bonding pressure, surface finishing, base material produced by different thermo-mechanical heat treatments, orientation of the base material, and the post weld heat treatment for secondary recrystallization. The optimum bonding condition was filtered and refined on the basis of the continuous grain growth structure across the welded interface, the size and the number of defects, the hardness distribution, and the mechanical properties of the joints. The fatigue strength of the joints thus optimized was also evaluated at high temperature in air, and was compared with that of the base material.
Axial strain-controlled fatigue tests were carried out at room temperature and 800°C for several kinds of TiAl intermetallic compounds with a duplex microstructure consisting of γ and lamellar phases. The results obtained in this study are as follows: (1) Fatigue lives of relatively ductile TiAl were predicted by Manson-Coffin equation using tensile ductility and slope of 0.5. (2) The fatigue strength at 103 cycles was correlated with the cyclic 0.2% proof stress. (3) The fatigue strength at 103 cycles was dependent on volume fraction of the γ phases.
In order to investigate the fracture process of a sintered silicon nitride at high temperatures, tensile, static creep, and fatigue tests were carried out at temperatures from room temperature up to 1573K, and the fracture surface of each specimen was carefully observed by means of optical and scanning electron microscopes. Fracture of the silicon nitride can be classified into two types from the viewpoint of cracking morphology. One is “single cracking type” where only a single crack brings about the fracture. As the crack is initiated at the largest pre-existing defect, distribution and size of the defects greatly affect the fracture strength. The other is “multiple cracking type” where several cracks were observed on the fracture surface as well as on the specimen surface. As the cracks are initiated by viscous sliding of glassy phase along grain boundaries, pre-existing defects affect little the fracture strength. The latter type is apt to appear in the tests with lower tensile stress or lower stress rate at higher temperature.
The effect of protective coatings on the high-cycle fatigue strength of Ni-base superalloy, IN738LC, was first modeled, summarizing the previous studies which covered the fatigue failure of the IN738LC specimens with three kinds of protective coatings: the CoCrAlY alloy overlay coated IN738LC, the CoNiCrAlY alloy overlay coated IN738LC, and the combined coatings in which the aluminide coating was conducted with different thickness on the CoNiCrAlY overlay coated IN738LC. The effect was found to be condensed into the following roles, depending on the mechanical properties of the coating material and substrate: the role as a fatigue crack initiator, the role to make a geometrical change of fatigue crack, and the role to induce a stress concentration. A new semi-empirical method to predict the fatigue endurance of the coated material was proposed, by applying the concept of latent, or fictitious crack, which has been originally introduced by Haddad et al. to explain the dependence of crack length on threshold of fatigue cracks in monolithic materials. In the proposed method the coating and the incidental diffusion zone were treated as a part of latent crack interacting with the substrate, depending on the mechanical properties of coating material. It was shown that the proposed method gave a good estimation corresponding well with the experimental results.