It is generally said that the compressive residual stress near the surface of tufftrided steel contributes to the improvement of fatigue strength. However, the mechanism of its contribution is hardly clear. In this paper, the author prepared several kinds of specimens for which the conditions of the compressive residual stress were different, and investigated the mechanism from a viewpoint of the fatigue crack initiation and propagation. The main results obtained are as follows; (1) The compressive residual stress had hardly an effect on the prevention of fatigue crack initiation and on the obstruction of fatigue crack propagation in the larger region of the stress intensity factor range, ΔK, too. However, the fatigue crack propagation in the smaller region of ΔK was delayed by the compressive residual stress. (2) Therefore, the compressive residual stress near the surface contributed to the improvement of fatigue strength in the case that the fatigue crack initiated at the surface of specimen, but it had little or no influence on the fatigue strength in the case that the fatigue crack initiated at the subsurface.
The mechanism of fatigue limit was examined on pure iron, which shows no strain ageing. The purpose of the experiments was to clarify whether the endurance limit is brought about by strain ageing or any other mechanism. The Jhonson-Matthey pure iron sheets were decarburized and denitrized in moist hydrogen for 70 days and 200 days. The former treatment gave the strain age hardenable sample and the latter the non-hardenable one. The latter sample showed also the fatigue limit at 107 cycles in pulsating fatigue tests at the stress level a little higher than the proportional limit. The age hardenable sample showed the knee in fatigue limit at 106 cycles at the stress a little below the tensile strength. The fatigue limit was concluded to be attributable mainly to strain ageing and hardly to work hardening. The conclusion above was applied to understand the problem of disappearence of the endurance limit under fluctuating fatigue stress on carbon steel. The technique here used was reversion treatment to remove strain ageing. The two S-N curves -one of which was obtained under constant applied stress and frequent reversion of strain ageing, the other was obtained under two-steps program load -coincided fairly good. The fatigue limit seems to exist only when strain ageing of the material is stable. If strain ageing condition is disturbed by occasional overstressing, the fatigue limit disappears and the S-N curve becomes similar to that suggested by Corten and Dolan. It is suggested that the fatigue limit is caused by balancing two mechanisms of hardening mainly due to strain ageing and damaging due to dislocation multiplication.
The behavior of fatigue cracks in plain specimens of eutectoid steel subjected to macroscopic tensile and shear stresses was examined by cyclic torsion tests at -60°C. The results obtained were compared with those at room temperature. The critical propagating stress of the fatigue crack was higher at -60°C than at room temperature for the same crack length. At the critical propagating stress, however, the size of slipped region at the crack tip and micro-Vickers hardness measured at room temperature in this region were little affected by both the crack length and the difference in fatigue test temperature. The hardness value in this region was smaller in the specimen subjected to tensile stress than to shear stress. The fatigue process from slip initiation to crack propagation at a crack tip was affected by the test temperature, the stress mode generated at the tip and the existence of the hardened layer at the tip.
The plane bending fatigue tests were made on a low carbon steel plate with a precrack under the stress ratio of 0.05 at room temperature. In some specimens, the precrack introduced by load cycle at room temperature was aged statically by heating in the short brittleness temperature range. By this procedure, the dislocation-relocking structures were introduced in the plastic zone left ahead of the precrack tip. In the other specimens, the precrack was introduced by load cycle in the temperature range above mentioned too. Therefore the plastic zone was aged cyclically. The crack propagation rate da/dn and the effective stress range ratio U were compared between the specimens aged and unaged. For the cyclically aged specimens, a decrease in da/dn was observed transiently. This behavior was similar to the well-known delayed retardation after a single high load. A decrease in U corresponding to that in da/dn was observed, too. The analysis based on the effective stress intensity factor range ΔKeff, however, showed that the crack closure concept could not perfectly account for the retardation effect in crack propagation. For the statically aged specimens, the retardation effect was observed too. The effect was smaller than that for the cyclically aged specimens. Both the statically and cyclically aged specimens showed the most remarkable effect on crack propagation behavior when aged at 375°C.
The conditions of the initiation of mode II fatigue crack growth have been investigated on the specimens with a fatigue precrack or a machined slit whose notch tip radius was 0.05mm or 0.5mm. Tests were made on 7075-T6 and 2017-T4 rolled plate and 2017-T4 extruded bar specimens. According to the test results, all the specimens of 7075-T6 and 2017-T4 with fatigue precracks and a 0.5mmR slit showed mode II growth under repeated mode II loading in the region of ΔKII larger than a certain critical value of ΔKII, where ΔKII of the crack of the same length is used for a slit with finite tip radius. The critical ΔKII's for mode II growth mentioned above depended on the material and on the slit tip radius. No mode II growth was observed on any of the 0.5mmR slit specimens of all the aluminum alloys tested. On 7075-T6 and 2017-T4 rolled plate specimens, tensile mode growth was observed in the region in which the ΔKII was smaller than that in the region in which mode II growth was observed. A fractographic investigation on mode II fatigue crack growth was made with special attention to the area of initial growth. In some cases similarities in fractographs of mode II growth and stage I growth in tensile mode growth were observed.
A new servo-controlled electrohydraulic fatigue testing apparatus which can operate in a field-emission scanning electron microscope (Hitachi HFS-2S) was designed. In order to allow direct observation of propagating cracks, piston and cylinder of the testing rig were connected by a link mechanism to move in counter directions so that loading could be applied simultaneously to the both ends of the specimen with arbitrary ratios. Moreover, some electrical means were used to obtain the stroboscopic-image of propagating cracks. Maximum loading capacity of the apparatus is ±5000N in tension and compression, with the maximum testing frequency of about 50Hz. Moving seal for high vacuum was made by using metallic bellows, and a turbo-molecular pump in addition to an ion pump was equipped for the vacuum system. Fatigue crack propagation tests were performed on a grain oriented 3% silicon iron at very slow propagation rates. In this range slips were found to occur at a relatively wide region near the crack tip, while in the high propagation rate range it had been reported that alternating slipping-off occurred only at the apex of the crack tip. So the different mechanisms of fatigue crack propagation were suggested for the high and the low ΔK regions for this material.
The fatigue crack growth resistance was evaluated for several aluminum alloys, and the factors to influence it were examined. The results were compared with the opening and growth characteristics of an ideal crack. The results obtained are as follows: (1) The ratio of the striation spacing to the stretched zone width for the same J value is independent of material properties and becomes nearly constant. By using this value, it is possible to predict the effect of fatigue crack closure. (2) The opening ratio, U, measured by ultrasonic technique is identical to that of the predicted value, although other conventional techniques have given higher values of U. (3) The main fatigue crack growth controlling factor is the flow stress, σfs, except 7075-T6 alloy, or Young's modulus, E, except 2XXX alloys.
Alternating two-load level fatigue tests were performed using annealed S40C steel to investigate the influence of overloads on the fatigue damage in crack initiation and propagation stages. The base block consisted of axial reversed strain cycling Δεp=1.1% as overloads and constant stress cycling σa=22kgf/mm2 as a baseline loading. Load levels were changed carefully so that no macroscopic residual strain could remain in the specimen. The load level at which the microcrack initiated in the first block determined the future crack growth behavior: In the case of large number of overloads in one block, microcracks initiated during overload cycling and the damage was cumulated linearly in the next base loading until cracks grew 0.1mm at the specimen surfaces. On the other hand, in the case of small number of overloads, microcracks initiated during next base load cycling and grew at higher rate than the normal; 0.1mm crack initiation life was reduced about 1/3 of the life for no overload and was about the same as that of the pre-strained materials. In the crack propagation stage, overloads caused crack growth retardation and a greater number of overloads resulted in a longer propagation life.
The results of a systematic investigation on the acceleration of fatigue crack propagation under periodic overstressing are summarized. The results show that a significant acceleration of fatigue crack propagation (amounting to more than one hundred times) occurs under variable stresses in which a very small number of overstress and a very large number of understress below the threshold stress intensity Kth are applied alternately. A method is proposed to predict the acceleration under complex multi-level variable stresses from test data under simpler two-level variable stresses. A reasonable agreement is obtained between the predicted acceleration and multi-level test data.
The impact fatigue tests were carried out by using a rotating disk type impact fatigue testing machine. The influence of impact loading, stress ratio and yield strength on fatigue crack growth rate was investigated by means of fracture mechanics and fractography in quenched and tempered 0.50% carbon steel. Then, the results of impact fatigue tests were compared with those of nonimpact fatigue tests. The results obtained are summarized as follows. (1) For the 423K tempered material, the dominant fracture appearance at impact fatigue was the dimple mode and the cleavage facet while that at non-impact fatigue was the intergranular facet. Such a fracture appearance was called the monotonic one. It was apparent that the acceleration of crack growth rate resulted from the rapid increase in the area percentage of monotonic fracture appearance regardless of the impact fatigue and the non-impact fatigue. (2) For the other tempered materials except the 423K tempered one, the striation was the dominant fracture appearance. However, the fracture appearance of the 573 and 873K tempered ones contained some inclusions and precipitation particles among the striation. (3) The effect of impact loading, Eac, defined by the ratio of impact fatigue striation spacing, Simp, to non-impact fatigue one, Snon, approached to about 1 as the yield strength, σys increased. The effect of impact loading could also be replaced by the equivalent stress ratio, Req. Then, the Req has related to σys as follows, Req=1-α·Eac-0.5=1-1.88×10-3·α·σys0.84. The Req decreased with increasing σys. This tendency was similar to the effect of stress ratio in non-impact fatigue. (4) The impact fatigue crack growth rate at a given ΔK level was larger than that in non-impact fatigue. This difference in crack growth rate increased with increasing tempering temperature. This fact could be well explained by the concept of equivalent stress ratio.
A statistical nature of fatigue life (total life) Nf was studied by decomposing Nf into the crack initiation life Ni and crack propagation life Np. By assuming that both Ni and Np follow lognormal distributions and that the coefficient of variation is larger for Ni than for Np, the distribution of Nf was derived theoretically by using Monte Carlo simulation. Mutual independence of Ni and Np was also assumed. It was shown that Nf followed a composite log-normal distribution approximately. In the regions left of and right of the transition point, this composite distribution was similar to that of Np and to that of Ni, respectively (the “transition point” denotes the point at which plot of Nf breaks into two segments on log-normal probability paper). The transition point moved downwards with an increase in the average ratio of Ni to Np. When this ratio was significantly small or large, Nf followed actually a simple log-normal distribution similar to the distribution of Np or to that of Ni, respectively. By combining the above results with the well-known experimental fact that the ratio of the crack initiation life to the crack propagation life increases with a decrease in stress level, a new interpretation was presented to the stress level dependence of the fatigue life distribution. This interpretation is different from the conventional one which is based on the mixed distribution. The present theory was examined experimentally. The data on an aluminum alloy and a low carbon steel showed that both Ni and Np followed log-normal distributions approximately and that the coefficient of variation was larger for Ni than for Np. Furthermore, the distribution of Nf as derived by Monte Carlo simulation was in good agreement with the measured distribution of Nf.
The role of mechanical factors in environmentally enhanced crack growth under cyclic loading was examined from a mechanistic view point in terms of time base crack growth rate in inert environment, [da/dt]inert, which was clarified to reflect the crack tip strain rate directly. Corrosion fatigue crack growth rate data in a simulated BWR environment at different stress ratios were compared with those obtained by the slow strain rate test (SSRT), aiming the direct correspondence on the growth rate diagram, [da/dt]E-[da/dt]inert diagram. This diagram can be used to develope the mechanistically-based design rule for environmentally enhanced crack growth such as corrosion fatigue, stress corrosion cracking and slow strain rate tests on the pressure boundary materials in LWR environments. Furthermore, it provides a physical basis to understand the acceleration mechanism and also a useful tool to predict the maximum acceleration under conceivable operating conditions. Finally, the usefulness of SSRT coupled with fracture mechanics is emphasized from its applicability to predictive testing and toughness evaluation in terms of JIc or tearing modulus Tmat in aggresive environments.
Cyclic crack growth behaviour of SA 533gr. B steel specimens of a common heat with different microstructures was examined through tests in a simulated BWR water environment. The variation in microstructure was provided by the following three different heat treatments, i. e. (1) air-quenched and tempered, (2) oil-quenched and tempered and (3) as oil-quenched, each of which was to simulate the localized variability of typical alternated microstructures of the weld-heat affected zones appearing along fusion lines. Crack growth rate measurements were made using constant-ΔK type contoured double cantilever beam (CDCB) specimens of 30mm thickness. The test environment, flowing 288°C 0.1-0.2ppmO2 water, was closely controlled by continuously refreshing the test section at a rate of 1000ml/min. In the preliminary air tests, no essential difference was observed among all kinds of materials in their ΔK versus da/dN relationships for two stress ratio conditions, i. e. R=0.1 and 0.5. In the BWR water environment, the crack growth rates of those three types of materials were accelerated relative to the air environment results. The degrees of acceleration were not greatly different to each other under stress ratio, R=0.1, while considerable difference was noted in case of higher stress ratio, R=0.5. Under the high stress ratio test condition, the quenched material, which was with martensitic structure and hardening, showed the highest degree of environment sensitivity among others. The results obtained were consistent with authors' earlier observation on the specific microstructure-dependent variability of the cyclic crack growth rates when cracks were advancing along weld fusion lines through HAZ, where a considerable level of residual stresses remained.
The fatigue crack initiation and the crack growth rate of chromium-molybdenum-vanadium steel in super heated steam were studied by the fracture mechanics approach in respect with the cyclic frequency and oxygen concentration in the surrounding atmosphere. Fatigue tests were conducted at 839K at cyclic frequencies of 600, 10, 1 and 0.1cpm. In order to examine the influence of environment upon the fatigue behaviors, the high temperature air and argon gas were also employed as the surrounding atmosphere. It was found that the number of cycles to crack initiation decreased remarkably and the crack growth rate increased with decreasing cyclic frequency and with increasing oxygen concentration. As for the influence of metallurgical factors upon fatigue behaviours, it was found that the number of cycles to crack initiation was remarkably reduced by the existence of inclusions such as manganese sulfide. The influence of environment upon the fatigue crack growth rate was examined under consideration of the fatigue crack growth rate per unit time.
It has often been reported that the hardening and compressive residual stress at surface were important factors for increasing fatigue strength of surface hardened materials. However, the effects of these two factors on fatigue lives were not always separated in previous experiments. Also it should be emphasized that no thorough investigation for these factors has been undertaken from a mechanical viewpoint. In the present study, the effects of laminated inhomogeneity and residual stress on bending fatigue lives were examined on the clad plates composed of low and medium carbon steels. On the basis of these experimental results, a method to estimate bending fatigue lives of the laminated inhomogenous metals was discussed. The main results obtained are as follows; (1) Cyclic bending stress-strain relation of the laminated inhomogeneous plates could be calculated approximately by applying the cyclic stress-strain relation of push-pull fatigue tests to the static bending deformation analysis. (2) Bending fatigue lives of the laminated inhomogeneous metals without residual stress could be estimated fairly well from S-N curves of the base metals by choosing the plastic strain amplitude (or strain amplitude) as an equivalent factor controlling the fatigue life. In the case with residual stress, the fatigue lives could be understood by adding an effect of residual stress on fatigue crack propagation lives after the microcrack initiation of fatigue to the estimation method mentioned above.
A series of fatigue tests had been carried out by using standard M9-10.9 bolts in order to show the influence of tightening bolts into yield by the angle-controlled tightening method on the fatigue strength of bolted joints. The tests were conducted with a 5 ton servo fatigue testing machine using a specially designed test fixture capable of subjecting alternating eccentric loads on bolted joints. As the initial bolt prestress was increased to yield, the additional bolt stress transmitted by the equal external alternating load was decreased. For this reason, the endurance limit of the bolted joints tightened to 90kg/mm2 became 15% above that of the bolts tightened to 60kg/mm2 and 60% above that of the bolts tightened to 20kg/mm2. On the other hand, the endurance limit of the bolt became less as the initial bolt prestress was increased. It could be interpreted through the fractographic examination using SEM that the above result was related to the influence of the initial bolt prestress into yield on the early stage of the fatigue crack propagation at the root of the thread groove.