Explosion hardened high manganese austenitic cast steel is being tried for rail crossing recently. From the previous studies, it became clear that high tensile residual stress was generated in the hardened surface layer by explosion and the surface tensile residual stress changed into the compressive stress of about 500MPa after the repetition of impact. In this study, the distribution of biaxial residual stress in the explosion hardened high manganese austenitic cast steel subjected to repeated impact loads was examined and compared with those of original and shot peened steels. The results obtained are summarized as follows. (1) The residual stress distributions in the impact surface layer of the original, explosion hardened and shot peened steels were similar qualitatively as given in the following. The surface compressive stress changes into tension at the depth of 0.01mm and the tensile stress in the surface layer shows a remarkable variation as the depth increases and changes into compressive stress in the inner layer. (2) It seems that the residual stress distribution mentioned above is produced by the combined effects of surface hardened layer, the transverse deformation in impact surface and uneven deformation at hardened layer. (3) The high tensile residual stress in the explosion hardened surface layer did not increase by repeated impact loading, and no macrocrack was osberved in the impact surface layer.
By applying Campbell's analytical method for the divergence condition of (∂ε/∂x) (ε; the true strain, x; Lagragian coordinate along pulling axis) to those of (∂e/∂x) and (∂A/∂x) (e; the nominal strain, A; the cross-sectional area), instability criteria in various tests were formulated for a rate sensitive material, and the correlation between the divergence condition of these imperfections and the load maximum condition was obtained in uniaxial tensile tests. The growth of the imperfection in the cross sectional area after necking was formulated in a rate sensitive as well as work hardening (softening) material (σ∝εm[1+α*ε], σ; the true stress, m and α*; constants). The analytical results give that not only m-value but α*-value have a significant effect on the stable growth of the area imperfection. The present analysis seems to be very useful to discuss such large elongation after necking as in superplastic material.
The understanding of stress relaxation behavior is important to evaluate the safety and reliability of structural materials used in high temperature plants. In this study, the long-term stress relaxation behavior was investigated for SUS 316 stainless steel at 500°C to 750°C. The residual stress gradually increased in the stress relaxation test below 550°C. On the other hand, the residual stress rapidly decreased after testing for 1000 to 3000hr above 600°C. The increase of residual stress at 550°C was caused by thermal densification due to the drop of carbon content in matrix accompanying the precipitation of carbide. The rapid decrease of residual stress at 650°C was caused by lowering of creep resistance due to the drop of Mo content in matrix accompanying the precipitation of Laves phase.
Effects of microstructures obtained by two kinds of thermomechanical treatment, i) solution treatment-cold rolling (67%)-aging and ii) solution treatment-aging-cold rolling (67%), on the fatigue behavior of an Al-1.8wt%Li alloy were investigated by means of optical and transmission electron microscopy. The results obtained are summarized as follows: Deformation bands perpendicular to the rolling direction were formed inhomogeniously in the cold-rolled specimens, and the dislocation cell structure and microbands were the characteristic features. Fatigue strength in both the thermomechanically-treated specimens was higher than those in the solution-treated specimens, in the solution-treated and cold-rolled specimens and in the aged specimens. The higher fatigue strength is due to the dislocation cell structure with δ'-Al3Li precipitates, because the cell structure alleviates the concentration of dislocation on a slip plane during fatigue test. Fatigue strength in the specimens with the deformation bands parallel to the axis of fatigue stress was higher than that in the specimens with deformation bands perpendicular to the axis of fatigue stress. The deformation bands perpendicular to the axis of fatigue stress became preferential nucleation sites and propagation path of fatigue cracks. The thermomechanical treatment, i) brought out higher fatigue strength than the treatment, ii).
The effects of austenite (γ) grain size, ageing at 475°C and volume fraction of γ (Vγ) on the fatigue strength and the process of fatigue crack initiation and early propagation have been investigated using (α+γ) two-phase stainless steels with different grain sizes including microduplex structures. The main results obtained are as follows; (1) Fatigue cracks were initiated within a γ grain and stopped at the α-γ interface at the fatigue limit for all specimens excepting the quenched ones with large volume fraction of γ. This means that non-propagating cracks are formed which have the same length as the γ grain size. (2) The relation between fatigue limit (σω) and γ grain size (dγ) could be expressed as follows for all the steels investigated. σωm·dγ=C The value of C in the equation for the steel aged at 475°C was larger than that for the quenched one. This result indicates that the fatigue limit of the steels can be improved remarkably by grain refinement and ageing at 475°C. (3) The fatigue limit for the specimens aged at 475°C depends on dγ only. On the other hand, the fatigue limit for the quenched ones depends on both dγ and Vγ.
Fatigue behavior of WC-12%Co hardmetal has been investigated under tension-compression cyclic loading at various stress ratios by using a newly developed apparatus. Tests were made under pulsating tension, pulsating compression and stress ratios R of -1, -4 and -9. WC-12%Co hardmetal showed two types of fatigue failure depending on the stress ratio. Under pulsating tension (R=0) and alternating stress (R=-1), fatigue cracks were initiated from the inside of the specimens. In these cases, cleavage fracture of Co was usually observed in the area of crack initiation. This type of fracture of Co is considered to occur in hcp Co transformed, due to the cyclic plastic straining, from the fcc Co which is the as-sintered state. In the fatigue crack initiation area, a Co-rich region or Co-rich regions were always found by EDX, in the specimens fatigue fractured under R=0 and -1. Fatigued specimens under pulsating compression showed a sharp decrease in tensile strength after a certain number of stress cycles. This decrease in tensile strength is caused by the grooves, which act as the fracture origin, formed on the specimen surface by scaling-off of surface layer. Fatigue fracture under tension-compression fatigue at R=-4 and -9 also occurred from the grooves formed on the specimen surfaces, just as the tensile fracture of fatigued specimens under pulsating compression did. In tension-compression (R=-4 and -9) fatigue, however, crack growth preceding unstable fracture occurred by fatigue due to the tensile stress component in cyclic stress, in addition to the groove formation by scaling-off of the surface layer.
Corrosion fatigue crack initiation behavior has been investigated in a high-tensile strength steel HT80 weldment in synthetic sea water. The tests were conducted on plane bending specimens. Constant amplitude sinusoidal loads were applied at a frequency of 0.17Hz with a load ratio of -1. In synthetic sea water, corrosion fatigue cracks were always initiated at corrosion pits. At high stress levels (σa≥147MPa), specimens were fractured at the softened area of heat affected zone, whereas final fracture at low stress levels (σa≤117MPa) occurred at the base metal. The fracture strength of HT80 weldments went down about 50% of that of HT80 plates in air or in synthetic sea water. The extreme value statistics were used to represent the growth of corrosion pits in the base metal, heat affected zone and weld metal at low stress levels. The pit growth rates in the base metal and heat affected zone were similar and were faster than that in the weld metal. By assuming the corrosion pits as sharp cracks, ΔK values at the deepest points were calculated and the mean value of ΔK, ΔK*CF, was obtained at 1.1MPa·m1/2. The proportion of life spent in crack initiation was estimated at about 40% of the total life at σa≤117MPa. Computer image processing technique was applied to corrosion pit analyses, and the 3-dimensional shapes of corrosion pits were obtained with a sufficient precision.
To clarify the effect of stress biaxiality on the fatigue crack growth rate, biaxial (torsional) and uniaxial fatigue crack propagation tests were carried out on a mild carbon steel. The torsional fatigue tests were performed using the tubular specimens with an initial slit 45°-inclined to the specimen axis. The behavior of fatigue crack closure was measured in each test by means of an unloading elastic compliance method. In the biaxial fatigue, the crack growth rate was faster under pulsating loading than under fully reversed loading, when the results were compared by using a parameter of usual cyclic stress intensity factor range ΔK. This difference in crack growth rate disappeared when the data were characterized in terms of a parameter of effective cyclic stress intensity factor range ΔKeff. Furthermore, it was found that the crack growth rate is faster under biaxial stressing than under uniaxial stressing.
In general, the studies of fatigue property of plastics have been carried out experimentally by rotary bending or plane bending fatigue tests and tension-compression fatigue tests in the axial direction, but hardly by torsional fatigue tests. In this article, the torsional fatigue properties of various engineering plastics (POM, PC and Epoxy resin) were investigated experimentally, and were compared with those by the rotary and bending fatigue tests. The results obtained are as follows: (1) The torsional fatigue failure of POM and PC, in which the static tensile strength σt is larger than the shear strength τ0, occurred by shearing stress in the first stage before the transition point B, at which the fracture mode changes from shearing to tension, and by tensile stress in the latter stage after the point B, whereas the Epoxy resin was always broken by tensile stress because its tensile strength is smaller than the shear strength. Therefore, there is no transition point B in the Sa-N curve of the Epoxy resin. (2) The torsional fatigue strength τf of POM, PC and Epoxy resin were 2.3, 0.9 and 1.7kgf/mm2, respectively, and the torsional fatigue strength ratios α(=τf/σt) became 0.34, 0.14 and 0.34, respectively, and the another ratios γ(=τt/τ0) become 0.38, 0.15 and 0.24, respectively. (3) The torsional fatigue strength of POM and PC were nearly equal to that of the plane bending, but were about 25-30% smaller than that of the rotary bending, and the torsional fatigue and rotary bending fatigue strengths of the Epoxy resin were quantitatively nearly equal.
The present authors have been investigating the delayed fracture process of soda-lime-silica glass in water by conducting four point bending tests using plate specimens with initial cracks and an indentation pit generated by Vickers diamond indentation. The crack growth behavior in the delayed fracture process, however, has not yet been clarified. Therefore, in this paper, the crack growth characteristics of the same glass meterials were investigated by employing the same plate specimens. At the same time, double cantilever beam test was conducted. Then, KI analysis of a semi-elliptical crack was applied to make some correction to a tentatively calculated KI value. In the analysis, the morphology effect of lateral crack was taken into consideration. As a result, the crack growth rate data obtained by four point bending tests were found to agree well with the double cantilever beam test data obtained in this experiment as well as those by S.M. Wiederhorn. Therefore, it is concluded that the KI-V relationship can be evaluated by the four point bending delayed fracture tests.
In this work, a new distribution function for fracture strength based on the multi-modal Weibull distribution function has been suggested in order to estimate the fracture behavior of brittle materials with ground surface. Furthermore, a new estimating method called “2-step estimation” for estimating the Weibull parameters accurately has been developed. The bending tests of alumina ceramics ground by a diamond wheel have been carried out and the following results were obtained. (1) The fracture origins of alumina specimens with ground surface were controlled by the competition between the pre-existing cracks and the machining induced scratch cracks. (2) The distribution function formulated in this work can explain the anisotropy of the bending strength of alumina specimens with ground surface fairly well. (3) The bending strength of alumina specimens used in this experiment was proportional to -1/2 power of the abrasive diameter in the region larger than 40μm.
Previously a stochastic model for damage accumulation was proposed to obtain an expression for the fatigue crack propagation rate. In this paper, the model is modified and the J-integral is introduced in order to derive the propagation law of creep cracks. On these theoretical bases, it is shown that the creep crack propagation rate is nearly proportional to the creep J-integral. Our theoretical results are compared with the experimental data.
An AC electrical potential system with a lock-in amplifier was developed to measure the propagation of a fatigue crack in notched specimens. A constant current of 1A with a frequency of 92Hz flowed into a single-edge notched specimen of a low-carbon steel. For a through-thickness fatigue crack longer than about 1mm, the potential difference was a quadratic function of the crack length. In an accurate measurement, the potential difference and the crack length at the end of the fatigue test were needed to determine the constant in the equation. An instantaneous resolution of this system was about 3μm, and a long-term error due to drift was about 10μm. When a crack was shorter than 1mm or part-through, two different quadratic equations were necessary to correlate the potential difference to the crack length. The system was successfully used to measure the crack length during the fatigue test of notched specimens.
Bromine (BRP) or chlorine modified Phenolic resins (CLP) were prepared by the reaction of paraharogenated phenol-phenol mixtures with formaldehyde. Boron modified Phenolic resins (PBP) were prepared by the esterification of boric acid with phenol, followed by the polycondensation with paraformaldehyde. The flame retardance of the modified resins was evaluated in comparison with the unmodified resin (GP) by the measurement of oxygen index, and found to be in the order: BRP>PBP>CLP> GP. The higher was the heat-treatment temperature, the higher was the heat distortion temperature (HDT) of resin. The HDT of resin was in the order: PBP, GP>BRP>CLP. This order was the same as the results of flexual properties. The resistance against thermooxidative degradation of the resins was examined by means of thermogravimetry and differential thermal analysis, and the higher resistance was observed by the order of PBP>GP≈CLP>>BRP.