The transmission shaft is one of the most important mechanical components and is widely used in machines. Usually, most of transmission shafts have steps and key ways, and fracture of those shafts is caused by crack initiation and propagation due to stress concentration from the key way. In practice, transmission shafts are commonly subjected to the combined stress from torsional and bending moments applied simultaneously. In consequence, the stress distribution is quite complex and it is difficult to analyze the stress appearing at the bottom of the key way. In the present study, stepped specimens with or without key ways were used and fatigue tests were carried out under combined stress of reversed torsion and reversed plan bending. From the test results, the following conclusions were obtained concerning the S-N curves, fatigue strength, the location of fracture and the initiation and propagation of cracks. (1) The effect of key way on fatigue strength increased in the order of bending, torsion and combined bending and torsion. This tendency was more remarkable as the fatigue life became longer. (2) Fracture on the surface of shafts with key ways was located in most cases at the end of the key way in the parallel portion of the specimen regardless of testing methods. (3) The location of crack initiation and the direction of crack propagation in the shafts with key ways subjected to fatigue tests under combined stress of bending and torsion showed an intermediate behaviour between those of bending and torsion. Cracks initiated in the plane of the principal stress, especially in the early stage of fatigue process.
In order to select a suitable specimen geometry for the evaluation of environment effect on fatigue property at HTGR temperatures, comparison was made between low-cycle fatigue strengths obtained for two types of specimens, solid cylindrical and hourglass specimens. A set of low-cycle fatigue tests were carried out on Ni-base wrought alloy Inconel 617 at 800°C and 1000°C in air and helium 4-nine purity. Strain was measured by two types of extensometers; (1) Hourglass specimen with a diametral extensometer, (2) Solid cylindrical specimen with a diametral extensometer, and (3) Solid cylindrical specimen with an axial extensometer. In all cases, the axial strain was controlled. The life of the hourglass specimens was consistently longer than that of the solid cylindrical specimens in both environments. The life of test type (2) was nearly equal to that of type (1) in air, but in helium the life of type (2) was between those (1) and (3). The effects of segregation and grain size on the specimen geometry were discussed. In the cases of the specimens machined at right angle to the segregation layer and those with fine grains, the difference of the lives obtained for two types of specimens increased. Furthermore, the difference of crack initiation behaviors in air and helium was discussed. The internal oxidation layer and the decarburized zone were observed near the specimen surface in helium, and more surface cracks occured in helium than in air. These may be related to the larger difference in lives of two types of specimens in helium, compared to in air.
A method of plastic analogy is proposed to construct a constitutive model describing the viscoplastic behavior of materials. By using the method, the knowledge obtained through the study of the plastic constitutive model can be utilized to construct the viscoplastic model. The plastic analogy is drawn because most of the viscoplastic models proposed previously can be written in the form of 3/2(Sij-αij)(Sij-αij)-κ2=0 and this is the same formula describing the yield surface of the plastic model except that the size κ of the surface is a function of the inelastic strain rate. The function κ can be determined from the data of the rate dependence of the yield stress obtained by constant strain rate tensile tests. As a demonstration of the method, a two-surface plastic model with a nonhardening strain region by Ohno is extended to a viscoplastic model. The viscoplastic model obtained here can be shown qualitatively to represent the plastic, creep and plastic-creep interaction behavior of materials with nonlinear strain hardening and cyclic strain hardening.
The structure of filament-wound composite is similar to that of unidirectional reinforcement materials. Therefore, two unique types of fracture appear in filament-wound composite under lateral load as in the case of unidirectional reinforcement materials. One of them is microscopic fracture which leads to a decrease in stiffness of filament-wound composite, and the other is macroscopic fracture due to interlamina delamination which causes large-scale dropping of lateral compressive load. Previously, the microscopic fracture had been investigated by using the elastic-plastic analysis. In this paper, the numerical model was improved in order to investigate the behaviours of both microscopic and macroscopic fracture. It was found that the numerical model considering the discontinuity of displacement at the node points and the Tresca's equivalent stress as the fracture criterion was in good agreement with the experiment results.
In the present study, observations of fracture process in aluminum were carried out by using a scanning exo-electron emission microscope. The distribution of the intensity of exo-electron emission near the notch root was also investigated. The intensity of exo-electron emission increased gradually with an increase of tensile strain up to the strain at muximum load. A significantly rapid increase in the intensity of exo-electron emission was found near the maximum load point, and then the intensity was saturated. From the results for the notched specimens, it was found that the crack initiation point coincided almost with the intersection of two tangent lines of the intensity of exo-electron emission versus displacement curve, one of which is in the region where the intensity increases rapidly and the other is in the region where the intensity is saturated.
Fracture toughness tests were carried out to clarify the effects of notch root radius, notch length and specimen thickness on the fracture toughness for transverse loading of unidirectional fiber reinforced plastics. As a result, it was found that an increase of notch root radius led to higher fracture toughness, while the fracture toughness was almost constant with the change of notch length or specimen thickness. In view of the observation that the fiber-matrix interfaces debonded first in the vicinity of the notch tip and then the macroscopic crack propagated partly through the debonding interfaces and partly through the resin, a model of crack initiation of the transverse cracking was proposed. By use of the model, the relation between fracture toughness and notch root radius was explained concisely. Furthermore, it was shown that the fracture toughness of the specimen with a crack, i.e. the minimum fracture toughness, was able to be estimated from that of the specimen with the notch.
Unlike homogeneous materials, composites have heterogeneous structure with distinct interfaces. These interfaces may be classified as microscopic surfaces within a lamina and macroscopic surfaces between laminae. An accurate evaluation of the stress distribution in the boundary layer should bring about a better understanding in the failure process of composites. Furthermore, in designing composite laminates, knowledge of the macroscopic interface stress is particularly important, since delamination of macroscopic adhesive surface results in a severe drop in strength. In this paper, an optimization procedure for the design of angle-ply laminate is offered, considering not only macroscopic behaviour as an in-plane problem but also non-uniformity of interlamina stress. At the beginning of the design process, in-plane modulus and interlamina stress were approximated as a constraint function and an objective function, respectively. And the ply orientation was optimized by non-linear programming (SUMT method). Consequently, the following results were obtained. A bidirectional laminate was optimized so as to minimize interlamina stress. As the design variables, the ply angle and the rigid-body rotation of a laminate from the principal axis were taken and they were determined by the optimization procedure. It is concluded that this procedure has wide applicability to various optimum design problems and is quite valuable in engineering.
In order to examine SCC of SUS 304 in industrial usage, the double bend-beam tests (four point bending) were carried out in low concentrated MgCl2 solutions under O2 flow, and the relation between SCC and crevice corrosion was examined. The results obtained are summarized as follow: In the double bend-beam test, cracks appeared on both of the sensitized SUS 304 and the solution-treated one under low MgCl2 concentration and low temperature conditions. The concentration and temperature in these experiments were low enough to be comparable to the SCC initiation condition in industries. The inter-crystalline cracks was observed on the sensitized SUS 304. Corrosion took place within the crevice between the double bend-beams. The initiation points of SCC were concentrated near the boundary between the crevice corrosion area and the non-corrosion area, where the passive state film seems to be very unstable. The crevice corrosion is considered to induce SCC initiation by concentration of chlorides and destabilization of passive state film in the crevice.
It was shown in the previous papers by the present authors that the fatigue crack initiation life of a notched specimen under axial load can be estimated from the Kσ-Kε relationship and σ-ε relationship of the material. In this paper, an attempt was made to apply this method to the case under bending or torsional loading by modifying the nominal stress of a notched specimen under axial loading. Fatigue tests were carried out under plane bending, and the results obtained coincided approximately with the estimated fatigue lives. In the case of rotating bending and torsional loading, the estimated fatigue lives were compared with the fatigue test data by other investigators and were found approximately to agree with them.
In the present investigation, the growth characteristics of small fatigue cracks were examined in two steels, and the critical crack length above which linear elastic fracture mechanics (LEFM) was applicable was evaluated. It is found that the critical crack length can be obtained as the sum of the microstructurally small crack length and the region of mechanically small cracks. The microstructurally small crack length and the region of the mechanically small cracks are also found to be an unique function of the microstructure (grain size) and yield strength of materials, respectively. Therefore, it is possible to evaluate the critical crack length for a given material.
In the general area of fatigue crack growth in the presence of residual stress, it is noted that the correction of stress intensity factor (K) to account for residual stress is important for the determination of both range and ratio (R) of K during a loading cycle. The superposition technique can be applied generally for the determination of K. For compact specimens, however, redistribution of residual stress occurs during crack growth and its effect is not negligible. In this study, experiments were done for the compact (CT) specimens which had such an orientation that the crack grew along the weld line in the weld metal of a butt weld joint of carbon steel. Fatigue crack growth characteristics of the weld joint were examined in comparison with that of the base metal and discussions for the evaluation of residual stress were made. The results obtained are as follows; (1) For the CT specimen, it is possible to predict the crack growth characteristics by using the initial residual stress distribution. (2) When the crack grows first through the compressive portion of a residual stress field, the effect is actually a result of the ratio (R) of K, although apparent high values of the crack opening stress intensity factor Kop were obtained experimentally. (3) The residual stress distribution through the thickness has no important effect on the average through-thickness crack growth rate, although it has some effect on the crack curvature. (4) Near the threshold region, the crack growth rate of the weld joint is higher than that of the base metal beyond the consideration of the residual stress effect.
An Experimental study has been carried out to determine fatigue strength and crack propagation behavior of press-fitted axles of a S45C carbon steel. Six kinds of specimens were prepared by means of heat treatments. That is, material A and B were normalized and annealed after heating at 850°C for 30min, C and D were normalized and annealed after heating at 1000°C for 4h, H was quenched and tempered after heating at 850°C for 30min, and R is as-received condition. For rotating bending fatigue tests, the press-fitted specimens, inserted two axles of 8mm in diameter into a joint with an interference of 10μm, were used to evaluate the fatigue strength, and the taper-fitted specimens, jointed the same axle with a tapered boss, were used to observe intermittently the crack propagation behavior. Although the fatigue strength of smooth specimens for all material conditions varied from 220 to 435MPa, that of the press-fitted specimen was in the narrow range from 140 to 190MPa. The fatigue strength of press-fitted axles was independent of the static strength or the plain fatigue strength of the material. From the observations of the crack propagation behavior at stress amplitude of 200MPa with the taper-fitted specimens, it was found that crack initiation was earlier in high strength materials, such as H and R, than in low strength materials, such as B and D. And, the crack propagation rate in the material R was higher than that in the material B or D, and that in the material H was lowest. Consequently, the fatigue strength of press-fitted axles was found to be lowest for the material R and highest for the material H.
A stochastic approach to fatigue crack propagation is proposed in consideration of random propagation resistence. It is based on the Paris-Erdogan's propagation law of fatigue crack. By adding the Gaussian white noise to propagation resistance of the propagation law, a non-linear Langevin equation is obtained. A stochastic differential equation of Ito type is derived from the Langevin equation by using a change of variable in the equation and Wong and Zakai's theorem. By using the solution of the stochastic differential equation and its probability density, a sample path and life distribution of fatigue crack propagation are derived, respectively. These theoretical results are compared with the experimental data for high tensile strength steel APFH 60. Through the comparison, an improvement in the above approach is made.
The statistical fatigue tests were carried out on S35C steel specimens by using three multi-type rotating bending machines designed previously. Fifty specimens were assigned to each of eight stress levels having the failure probability from zero to 100 percent. Based on the experimental results, Weibull three parameters were estimated and their stress dependency was investigated. The results are summarized as follows: (1) It was observed that some specimens survived 108 cycles at the stress levels below σ=31kgf/mm2 (304MPa), and in this stress region the scatter of lives became very large. An irregularity was observed in the same region that the shortest life of 50 specimens became small with a decrease of stress. (2) Based on the Weibull three parameters estimated from the correlation coefficient and the saturated failure probability, the experimental fatigue life distribution was well fitted by the Weibull distribution. Regarding the stress dependency of the parameters of Weibull distribution, the shape parametar A took a value of 0.7-1.0 in the region of σ≤31kgf/mm2 (304MPa), and it increased gradually with an increase of stress in the region of σ≥32kgf/mm2 (314MPa), but no acceptable relationship was found between the stress and the scale or location parameter. (3) Since the normal distribution did not agree well with the experimental results of the endurance limit distribution, at its extreme regions, an attempt was made to apply a beta distribution defined in the finite region. Good agreement was obtained throughout the whole range of experimental data.
In order to study the effects of vacuum environment and strain wave form on high temperature low-cycle fatigue properties of Hastelloy XR, a set of strain controlled fatigue tests were conducted with sawtooth type and strain hold type strain wave forms at 950°C in air and vacuum environments. The following conclusions were obtained; (1) The effect of environment on life under symmetrical wave forms is as follows. The lives in vacuum are longer than those in air. The environmental dependence of fatigue life can be explained based on the properties of high temperature oxide scale and oxidation behaviors. (2) The fatigue lives under sawtooth type wave forms decreased as follows; Fast-Fast>Fast-Slow≥Slow-Slow>Slow-Fast, and the life under strain hold with the same period decreased as follows; hold in compression>hold in both sides>hold in tension. The fatigue lives under symmetrical strain wave forms were affected more by the environment than that under asymmetrical waves. (3) Intergranular failures dominated in both of the tested environments. The difference in environmental effect between two loading wave forms can be explained based on the damage model for unidirectionally accumurated strain due to grain boundary sliding.
The combined creep-fatigue loading tests on 1Cr-Mo-V steel were carried out under the creep loading conditions in the transgranular fracture region on the creep fracture mode map obtained from the static creep rupture tests. The data obtained were evaluated using the linear life fraction damage rule. The specimens under combined creep-fatigue loading were ruptured by an accumulation of either creep damage (φc) or fatigue damage (φf), and the values of φc and φf at fracture were about 0.5 and 0.8, respectively. This relation for 1Cr-Mo-V steel showed the same trend as that for SUS 304 or 316 stainless steel obtained from the combined creep-fatigue loading test with the same loading condition, but the values of φc and φf for 1Cr-Mo-V steel were different from those for the stainless steels. It was considered that this difference was mainly due to lowering of creep rupture life arising from softening of 1Cr-Mo-V steel during loading. The φc vs. φf relation obtained for the present steel almost agreed with that obtained from the interspersion test in MPC.
An investigation of stable laser action at room temperature using the F2 centers of LiF: Mg crystals is presented. A remarkable reduction in fading of the output signal intensity has been achieved by adjusting the Mg impurity level to an appropriate quantity and by optimizing F2 center concentration. It is found that magnesium ions in LiF crystals play a role for suppressing the fading of laser action due to the two-step photoionization process. It is also found that the fading strongly depends on intracavity power and temperature.
Many studies on arc deterioration of electrical insulating materials by high voltage-low current arc have hitherto been done by using small size samples with electrodes having short gap length. In this paper, the arc resistance tests different from the previous ones were carried out by applying higher voltage and longer electrode gap length than before. Arc resistance was measured by a new arc resistance test method and the process of arc deterioration was investigated. As the experimental method, ASTM D495 standard test method was modified a little. Electrodes were the stainless steel plate type of 5-100mm in length and the applied voltage was max. 100kV. Furthermore, arc resistance in the case of electrodes detouched from the sample surface was also investigated. The experimental results are summarized as follows: (1) The arc resistance of insulating materials measured with contact electrodes having long gap length showed two different tendencies depending upon the species of insulating materials, one showing a slow increase with increasing gap length from about 40mm to 100mm, and the other a tendency of saturation at the gap length of 50mm or more. Thus, the effect of electrode gap length on arc resistance was observed clearly. (2) The arc resistance of insulating materials measured with detouched electrodes having long gap length also showed two different tendencies depending upon the species of insulating materials, one showing a slow increase with increasing gap length from about 30-40mm to 100mm and the other nearly constant. Thus, the separation of electrodes from the sample surface affected arc resistance remarkably and the effect of electrode gap length on arc resistance was observed in this case also. (3) The effect of spatial configuration of sample on the arc resistance of insulating materials exists not only in comparatively small electrode gap length but also in long one.