In construction works, bolts used for connecting structural components are often subjected to severe conditions. These bolts are not always made of high toughness materials. Therefore, collapses of the components are sometimes caused by failure of the bolts. In this study, to evaluate the strength of these bolts, tensile and fatigue tests were carried out on two kinds of bolts employed to build up caisson shafts. One of them was a group of the bolts which had been actually used in caisson works and the other was the unused one. Then by assuming a simplified model for a caisson shaft, stress in the bolt needed to connect blocks of the shaft was estimated by calculation and the results were compared with the experimental data. A large amount of scatter was found in both tensile and fatigue strengths for the used bolts, as compared with those of the unused ones. It was also found that the maximum stress obtained by calculation reached almost to the yield strength of the tested bolts and the calculated stress amplitude was greater than their endurance limit. From these results, the followings are recommended to prevent failure of the bolts for caisson construction: (1) to employ the bolts with higher strength than those which are generally accepted at present and (2) not to use the bolts which have been used once.
The stress applied to a specimen does not fully act to deform material. The effective stress which actually acts on dislocations can be related to the internal stress of material. The origin of internal stress and thus effective stress has been discussed here by means of a dislocation model, concluding that internal stress depends on the dislocation structure. Based on this conclusion, fatigue tests were carried out and the following results were obtained: (1) The behavior of internal stress in the fatigue process showed a clear difference between the over-stressed and the under-stressed conditions. (2) The fatigue failure occurred under many-fold multiple repeated stress in two stress levels below the fatigue limit. Such fracture was attributed to the instability of internal stress. (3) The concept of internal stress and effective stress is useful to the comprehensive understanding of fatigue phenomenon.
Fatigue strength data of notched specimens of steels were analyzed by distinguishing the fatigue limit for crack initiation, σω1, and that for fracture, σω2, with the aid of fracture mechanics. The data were extracted from Data Book on Fatigue Strength of Metallic Materials published recently by Japan Society of Materials Science and JSME Data Book, entitled Design Data on Fatigue Strength of Metals (I)published by Japan Society of Mechanical Engineers about 20 years ago. Although there existed little difference in characteristics of fatigue strength data between the above-mentioned two data books, empirical formulation of fatigue strength reduction factor, β, proposed by JSME was found to be unsatisfactory in the sense that σω1 and σω2 were not distinguished and to give an unconservative estimation for the size effect on σω2 as a design criterion. As for the σω1, since machining conditions of test specimens might give a large effect on fatigue strength, it seemed difficult to derive better information from this kind of data books other than the one that the stress concentration factor, α, provides the upper limit of β1. As for the σω2, the data are urgently needed for large sized specimens to elucidate the size effect on σω2 and to discuss the possibility of applying fracture mechanics to fatigue of notched specimens.
The fatigue crack initiation life prediction of butt welded joints with notch under pulsating compression (R=-∞) by bending load has been investigated by paying attention to the local strain at the notch root. The experimental results indicate that the fatigue crack initiation and propagation take place at the notch root in spite of pulsating compression, and there is a tendency that the propagation rate decreases with increasing cycles. These cracks emanating from the notch root under pulsating compression become nonpropagation cracks. From an engineering point of view, the fatigue crack initiation life at the notch root can be estimated by a local strain approach using the fatigue crack initiation data of smooth specimens. It becomes clear that in consideration of the compressive plastic-induced opening the fatigue crack growth behavior at the notch root under pulsating compression can be well explained by the effective stress intensity factor.
Push-pull fatigue tests under different mean stresses (σm=-98, 0, +98MPa) were carried out on annealed low carbon steel containing an artificial drilled hole (the diameter of the hole: d=50, 100, 200μm). The experimental results were compared with those of rotating-bending tests. The geometrical parameter controlling ΔKth was proposed on the basis of the analytical results on surface cracks with various shapes. The conclusions are summarized as follows: (1) The effect of defects increased with increasing mean stress from compressive to tensile stress. This is because the propagation of cracks emanated from defects is dependent on the magnitude of mean stress. (2) The true stress at the surface of rotating-bending fatigue specimens was calculated on the basis of cyclic stress-strain curve and the true stress of fatigue limit under bending was found to be nearly equal to that of push-pull fatigue under zero mean stress. (3) ΔKth of the specimens containing small surface cracks with arbitrary shape can be evaluated uniquely using the geometrical parameter √area, where area is the area of a defect projected in the direction of the maximum principal stress. The effect of mean stress on the relationship between ΔKth and √area was also shown.
The characteristics of small fatigue crack growth were investigated under rotating bending on two specimens of a low alloy steel having prior austenite grain sizes of d=15μm (fine grain) and d=91μm (coarse grain). The influences of grain boundary on the crack growth rate and the aspect ratio of small fatigue cracks were clarified, and the critical crack length above which linear elastic fracture mechanics (LEFM) was applicable, was evaluated for growing small fatigue cracks. When the surface crack length was shorter than 3d, the crack growth rate decreased near grain boundary, and the aspect ratio varied widely by the effect of microstructure. Cracks longer than 3d had no influence of the microstructure, but they grew faster than would be expected based on LEFM until the length reached to 3d+150μm. This behaviour may be attributed to the difference in crack closure between small cracks and large ones. If the contribution of the crack closure to the small fatigue crack growth were to be established experimentally or analytically, the critical crack length would be 3d. However, it is now difficult to evaluate the crack closure and therefore 3d+150μm is regarded as the critical crack length for engineering application at present.
It has been considered that the crack propagation rate is controlled by the tensile part of stress range but not by the compressive part. However, according to the experimental results carried out by the present authors for a plane bending stress condition, the compressive part of stress range affected the crack propagation behavior and the accelerated crack propagation rate was observed, causing shorter fatigue life than that without a compressive part. To clarify the effect of the compressive stress range of cyclic amplitude on the crack propagation, push-pull fatigue tests with a constant stress amplitude two step multi-fold stresses were carried out for plane sheet specimens (CCT) of low carbon steel (the ultimate tensile strength is 517MPa). The crack opening stress level was measured by the unloading elastic compliance method. The conclusions obtained are summarized as follows: (1) When the crack propagation rate was high, the compressive part of stress range accelerated the propagation, due to the lower crack opening stress level. (2) When the compressive part of stress range, that is the absolute value of minimum stress in cyclic stress range, increased beyond a certain value, the acceleration of crack propagation rate appeared. The compressive part of stress range larger than this value showed the same acceleration effect on the crack propagation behavior. (3) The acceleration effect of the compressive part of stress range on the crack propagation rate was observed in the tests where the stress ratio R1 in one step was 0 and R2 in another step was -0.33. (4) When the crack propagation rate was low, the comppressive part of cyclic stress range decelerated the crack propagation rate because of the oxide induced crack closure.
The behavior of mode II fatigue crack growth and its mechanism have been investigated on 2017-T3, 7075-T6, 7075-T651 aluminum alloys and 7N01-T4 aluminum alloy weldments by using a four-point-shear-loading method. In the aluminium alloys tested, the da/dN-ΔKII relation and ΔKII-threshold for mode II fatigue crack growth were found to be very material-sensitive, while da/dN-ΔKIeff relation for mode I growth was rather material-insensitive as is well known. Observation of the deformation around a growing crack tip has been made on the repricas taken at the successive stages-minimum load, maximum load, minimum load-in one load cycle. Two types of mechanisms of mode II fatigue crack growth have been proposed based on the observed results. In 2017-T3 and 7N01-T4 base metal, HAZ and weld metal, a mode II fatigue crack grows due to the alternating shear slip mechanism at the crack tip, and in 7075-T6 shear-induced-decohesion takes place in addition to the alternating shear slip. The cyclic crack tip shear displacement Δ(CTSD)p on the boundary of the cyclic plastic region measured on the repricas agreed well with Δ(CTSD) calculated from BCS theory in all the materials tested. It has been confirmed that the fractographical features observed support the proposed mechanisms of mode II fatigue crack growth.
In the case of two-level cyclic stress loading, it is generally recognized in steels that the cumulative cycle ratio (ΣN/Nf) is greater than unity for a low to high stress sequence (σL→σH) but less than unity for a high to low stress sequence (σH→σL). This kind of deviation from the Palmgren-Miner's rule (linear damage rule) is expected to be related to the fatigue damage accumulation process and age-hardening ability. Two kinds of low carbon steels were prepared; specimen A, in which the age-hardening ability was attained through low temperature quenching, and specimen B, in which the ageing process in specimen A was completed by leaving them at room temperature. In specimen B, the deviation from Palmgren-Miner's rule was a little for both σL→σH and σH→σL sequences. On the other hand, in specimen A which had age-hardening ability, the deviation from Palmgren-Miner's rule for σL→σH sequence was greater than that in specimen B. The change of internal friction energy and propagation of fatigue crack were observed during cyclic loading. Based on the experimental results of accumulation process of internal friction energy, the following has been assumed. (1) A part of the internal friction energy is dissipated for the accumulation of fatigue damage. (2) Crack growth corresponds to fatigue damage. (3) The accmulation process of fatigue damage is expressed by the Marco-Starkey and New-mark models in which an exponent depends on stress amplitude and increases with progress of ageing. Based on the above assumptions, the cumulative cycle ratio was calculated, and the result showed a similar tendency as the experimental one.
The constant and variable stress amplitude fatigue tests were carried out on smooth and pinholed specimens of SCM 435 steel under rotating bending conditions. In the constant stress amplitude tests, the fatigue crack propagation rate was not affected by the stress level and it was represented well by ΔK. The crack initiated by the over-stress σa1 cycles was retarded by the under-stress σa2 cycles, when the difference between σa1 and σa2 was large. In the case of non-propagating cracks, their length was shorter than that in plain carbon steels. The condition of crack retardation under σa1 repetition following σa2 application was elucidated. The crack propagation behavior of pin-holed specimens in two step multi-fold tests was similar to that in two step two-or three-fold tests. The discussion was made on the shift from unity in cumulative cycle ratio for smooth specimens, based on the results of crack propagation behavior of pin-holed specimens.
Fatigue crack growth tests were carried out on two aluminum alloys, ZK141-T7 and A7075-T6, under constant amplitude, repeated two-step and random loadings. Crack length and crack closure were measured by using the minicomputer-aided unloading elastic compliance method. The crack growth rate curve represented in terms of effective stress intensity range, ΔKeff, exhibited a trilinear form in so-called Region II for both testing materials under constant amplitude loading. The relation between da/dn and ΔKeff of ZK141-T7 under repeated two-step loadings was found different from the constant amplitude results due to load interaction, whereas the relation of A7075-T6 was hardly affected by load variation. The crack opening stress intensity, Kop, under variable loadings was found for both materials to be controlled by the maximum stress intensity range-pair, (ΔrpK)max, and its stress ratio, and further to coincide with the constant amplitude test results having the identical stress intensity range ΔK and stress ratio R. Fatigue crack growth under random loadings could be well predicted in terms of ΔKeff, using the modified relation between da/dn and ΔKeff obtained from the repeated two-step loading tests accounting for crack growth characteristics depending on materials.
Nominal stress-controlled low-cycle fatigue life under imposed mean stress is affected by elongation or contraction due to cyclic stress-induced creep. Therefore, the effect of mean stress was examined by a microcomputer-assisted testing device which provides true stress controlled loading. Through investigation of a medium carbon steel (S45C), the following results, were obtained: (1) Cyclic plastic strain amplitude at the steady cyclic stress-strain state decreased with an increase of the imposed tensile mean stress. (2) The cyclic stress-induced creep proceeded so far as the mean stress was imposed. The increase in axial cyclic creep strain per cycle was proportional to the nth power of the maximum cyclic stress. (3) For the same cyclic stress amplitude, the application of tensile or compressive mean stress prolonged the fatigue life as compared with the fully reversed stressing. (4) A modified Manson-Coffin relation including such two parameters as the cyclic stress amplitude and the mean stress was proposed for evaluating the fatigue strength under cyclic stress-induced creep condition.
A cryostat has been developed for this study, which enables to conduct axial-strain controlled fatigue tests of SUS 304L stainless steel at 4K with the strain range between 1.04 and 2.52%, the strain rate of 0.4%/sec and the strain ratio of -1. The loading in this cryostat is achieved by inner and outer tubes so that the requirement for high structural rigidity to suppress the specimen buckling and that for the low cross section area to keep the heat leak as low as the liquid He consumption rate of 1.5l/H or less are both satisfied. The fatigue tests at 300 and 77K were also conducted for comparison. On the curves showing the changes of maximum and minimum stresses with the number of cycles as well as the stress-strain curve, the serration-like stress change was observed at 4K, while the smooth stress change was observed at 300 and 77K. The cyclic stress-strain curves at 4 and 77K intersected each other, showing that the stress is lower at 4K than at 77K in a small strain amplitude. Those curves at 4 and 77K lied far away above the one at 300K. The fatigue life curves or the total strain range vs. fatigue life curves at 4, 77 and 300K also intersected each other, having a trend that the fatigue strength at 4K is the highest in the high cycle range but becomes the lowest in the low cycle range. The fatigue life curves of SUS 304L obtained here at 4, 77 and 300K gave almost the same fatigue strength as that of AISI 304L obtained under the diameter-strain control condition. When compared with other austenitic stainless steeels such as AISI 304, AISI 316 and AISI 310, the fatigue strength of SUS 304L was lowest at all temperatures.
In order to investigate the corrosion effect on the fatigue strength of steel sheets used in automobile suspension members, plane bending fatigue tests were carried out. The materials used were 0.075%C and 0.15%C hot-rolled steel sheets. One group of specimens were intermittently dipped in 5% salt-water at room temperature and another group continuously in 20% HCl aqueous solution at 50°C. The fatigue strength of the corroded specimens was correlated with the feature of the corroded surfaces. The main factors affecting the degradation of fatigue strength due to corrosion were the reduction of the sectional area and the increase of surface roughness. The corrosion products and charged hydrogen did not decrease the fatigue strength of the corroded specimen in a dry condition. The following equation was proposed for estimation of the bending fatigue strength σwap of the specimens after the corrosion products were removed. σwap=(Zc/Z0)·(1-ηΔdmax0.4)·σw0 where Z: section modulus, η: 0.45 for 0.075C steel and 0.65 for 0.15 steel, Δdmax: maximum depth of roughness due to corrosion, suffix 0: Electropolished specimens, suffix C: Corroded specimens.
Two-stage corrosion-fatigue life tests in 3% NaCl_??_laboratory air, 3%NaCl_??_ion exchanged water, and 3%NaCl_??_0.5%NaCl conditions were performed under torsional load, and the results were discussed based on the observation of cracks on the axial section of test pieces. The following conclusions were drawn: (1) For the test of 3%NaCl→laboratory air, the fatigue life was reduced remarkably and the total number of stress cycles to failure became shorter than that for the continuous 3%NaCl condition. A similar tendency was obtained for the tests of 3%NaCl→ion exchanged water, and 3%NaCl→0.5%NaCl. However, the reduction of fatigue life was slight because the damage in the first stage was not so severe as that for the test of 3%NaCl→laboratory air. (2) For the reverse tests, i.e. laboratory air→3%NaCl, ion exchanged water→3%NaCl, and 0.5%NaCl→3%NaCl, the fatigue lives followed Miner's law. (3) After the corrosion environment was changed to less corrosive one, a few cracks among many short cracks having appeared in the first stage grew acceleratedly and led to a shorter life. For the reverse case, most of the cracks propagated in the second stage, and the life was predictable by the Miner's law.
Fractographical observations of low alloy steels in high temperature water were carried out, and the responsible mechanisms for corrosion fatigue crack growth were discussed. The fracture surface of a low sulfur content steel was wholly occupied by transgranular ductile cracking characterized by striation patterns. In a medium sulfur content steel, on the other hand, quasi-cleavage cracking characterized by both striation and river patterns was observed. The crack growth in the medium sulfur content steel was abruptly accelerated at ΔK≅19MPa·m1/2. This acceleration was caused by hydrogen embrittlement associated with dissolution of MnS. Fourier power spectrum analysis was found to be a very powerful tool for the striation spacing measurement, especially when the striation patterns were made indistinct by derusting treatment. Based on the linear summation hypothesis with consideration for fracture area fractions, the quasi-cleavage crack growth rate in the medium sulfur content steel was derived.