The stress intensity factors of semi-elliptical surface cracks in a finite thickness plate subjected to bending were estimated by taking account of the front free surface magnification factor for stress intensity component due to linearly varying pressure, which was neglected in the well-known calculation by Shah and Kobayashi. The stress intensity factors obtained were somewhat smaller than those given by Shah and Kobayashi, and gave a reasonable estimate of the propagation rate of part-through cracks by fatigue, in the case where the crack length in the depth direction was less than a half of the plate thickness.
As a preliminary examination for evaluating the effect of combinations of several varying environmental factors on fafigue cumulative cycle ratio Σ(n/N), the plain bending fatigue tests were carried out with various combinations of stress amplitude and temperature. The combinations used were: (1) a constant stress amplitude and step-up-2 stage or step-down-2 stage temperatures, (2) a constant temperature and step-up-2 stage or step-down-2 stage stress amplitudes, (3) a constant stress amplitude and multiple repetition of 2 stage temperatures, (4) a constant temperature and multiple repetition of 2 stage stress amplitudes, and (5) multiple repetition of 2 stage temperatures and 2 stage stress amplitudes. The following conclusions have been obtained. (1) Σ(n/N) became larger than unity under the tests for combinations of the step-up temperature environment and constant stress amplitude, and of the constant temperature and step-up stress amplitude. (2) In the completely opposite cases to the conditions discribed-above, the values of Σ(n/N) seem to be smaller than unity. (3) In the cases of the multiple step-down-2 stage for the both factors (500°C_??_300°C, 32kg/mm2_??_28kg/mm2), if the period of stage for the stress amplitude 28kg/mm2 at 300°C is short, the values of Σ(n/N) are smaller than unity but, if this period is long, the values of Σ(n/N) are extremely larger than unity. (4) The scatter band of the cumulative cycle ratio for all the test results was in the range of about 0.2 to 100. Moreover a new idea for predicting fatigue life of machine parts employed under various combinations of temperature, corrosion environments and stress amplitude is proposed on the basis of the knowledge obtained previously from the tests under various combinations of corrosion environment and stress amplitude and the present informations.
The arc resistance of electrical insulating materials is an important subject in the insulation design of electrical equipment. At present, ASTM D495-71 test has been designated as the standard arc resistance test method to simulate the service conditions. However, it is considered that a complete elucidation of their properties at the failure of electrical insulation against various types of electrical arc is impossible by only one test method. Consequently, the incorporation of other test methods such as a modified ASTM arc resistance test with the standard method is desirable to the proper appraisal of arc resistance of plastics. In this paper, various test methods such as the rotary disc arc resistance test, the middle current arc resistance test, the high frequency arc resistance test, the electrical furnace arc resistance test and the arc resistance test by the chemical flame were used for the measurement of arc resistance. As the test samples, various plastic materials were used. The mutual comparison of experimental results showed that the combination of various measured values is useful to the appraisal of arc resistance. Among the test plastics, the resins such as silicone glass and melamine glass show large arc resistance and the resins such as phenol resin, epoxy paper-filled, tetron epoxy laminate insulation and epoxy glass cloth show comparatively small arc resistance.
Kinetics of phase transformations in steel during quenching is presented. The transformations form austenite to pearlite and to martensite are characterized as a function of thermal history, and the volumetric change of an element is discussed. A plastic flow rule including such effects of structural changes is proposed in the second part of this paper. A numerical example is shown for a semi-infinite body cooled at the surface from an eutectoid temperature.
Elastic-plastic stress and strain concentration factors under tensile load have been examined for notched sheet, notched rectangular and circumferentially notched cylindrical specimens. For the notched sheet specimen the feature of variation of elastic-plastic stress and strain concentration factors obtained experimentally is similar to that calculated by the finite element method. The elastic stress distribution measured by using strain gages along the notch root of the rectangular specimen notched at two sides shows that both longitudinal and transverse stresses become maximum at the center of the notch root and decrease with approaching the end of the notch root. The state of stress at the notch root is biaxial tension. The elastic-plastic strain concentration factors computed by the finite element method for the rectangular specimen in plane strain state and the cylindrical one in axisymmetrical deformation increase rapidly with plastic deformation of the notch root but decrease rapidly to less than unity after the maximum value.
A critical examination was made on the applicability of fracture mechanics parameters to crack propagation under creep condition in 304 stainless steels with a wide variety of specimen geometries at a high stress level. The creep crack propagation rate could not be described in terms of elastic stress intensity factor but it could be in terms of net section stress for all dimensions of thin plate specimens. The net section stress, however, could net explain the difference between the crack propagation rates of thin plate specimens and those of notched round bar specimens. This difference could be ascribed to the plastic constraint around the crack tip. Applicability of the modified J integral to creep crack propagation was also examined and the results showed that this parameter was the best one to predict the creep crack propagation rate at high stress level. In the case of specimens of similar geometries, the net section stress could be used in place of the modified J integral provided that the crack propagation rate was divided by the characteristic length of the specimen.
Experiments were conducted on the propagation of part-through cracks (semi-elliptical surface cracks) by alternating bending stresses with sheet specimens of an aluminum alloy (2017-T3). Crack closure levels were determined by means of the compliance method as well as by the surface observation. The propagation rate of part-through cracks could not be correlated with that of through-thickness cracks in terms of the stress intensity, but it could be in terms of the effective stress intensity based on closure measurements by the compliance method. The effective stress intensity based on closure measurements by the surface observation was applicable neither to crack propagation in the interior nor to that on the surface of specimens.
In the previous paper, the relation between fiber/matrix and FRP/FRP interfaces and its dependence on matrix were discussed on the basis of the tensile tests of FRP along its flatwise direction. The results obtained showed that the location of the weakest interface varies depending upon the condition used and so the matrix dependent property of interface may be evaluated by such a test. In this paper, the relation among three interfaces (fiber/matrix, matrix/matrix and FRP/FRP interfaces) and its dependence on matrix were studied by using roving glass cloth FRP with the matrix/matrix interface formed after various resting times before the wet on dry laminating process. The results obtained are as follows. (1) In the case of FRP/FRP interface with sufficiently high bond strength, the weakest interface is fiber/matrix interface irrespective of matrix or laminating condition in a nomal state. Therefore, the bond strength seems to vary little and the difference in strength due to the thickness effect produced by the condition of wet on dry laminating process depends on matrix but it does not essentially affect the fracture mode. (2) The weakest interface under the condition of immersion in boiling water depends on matrix, laminating condition and immersion time. Especially the location of the weakest interface is affected by the thickness effect in addition to the resting time before the wet on dry laminating process and the immersion time, and it depends also on the property of matrix. (3) In general, fiber/matrix interface is the weakest among the three but either FRP/FRP or matrix/matrix interface becomes the weakest interface in some other conditions. Therefore, by selecting such conditions, it is possible to evaluate the property of interface in FRP individually.
This study was performed to develop a high reactive epoxy resin system which can be used without heating in winter construction works in cold northern districts. Firstly, the following 3 experiments were conducted. (1) The effect of curing accelerators (66 kinds) on the gelation time of an epoxy resin. (2) The variation in viscosity of the mixtures (24 kinds) of polyepoxide (or hardener)-curing accelerators with a lapse of time. (3) The temperature dependence of viscosity of the mixtures (12 kinds) of polyepoxide (or hardener)-curing accelerators. As a result, it was found that a mixture of hardener-m-cresol has the best quality for practical use with regard to low viscosity, low increase rate of viscosity with lapse of time, low lower limit of usable temperature, and high reactivity. Secondly, 8 kinds of test materials (mixtures of hardener-m-cresol) which have the m-cresol contents ranging from 0 to 140 equivalent per cent at an interval of 20 equivalent per cent were made, and were used for the following 2 experiments. (4) The temperature dependence of viscosity of each test materials. (5) The temperature dependence of gelation time of the mixtures of each test material-polyepoxide. The results obtained from these two experiments were summarized as follows: (1) The test materials which have the m-cresol content below 120 equivalent per cent can be used at the average temperature in January in all cities in Hokkaido with 1 exception. (2) The test materials which have the m-cresol content above 80 equivalent per cent can be used below the freezing point.
To decide pV-limit applicable to a wide range of experimental conditions, the friction and wear tests are carried out by using ultra high molecular weight polyethylene (UHMWPE), CFRUHMWPE and γ ray irradiated ones as samples. Two distinguishable phenomena in friction and wear have been observed. One is a usual phenomenon; the frictional force and the temperature of friction surface increase as the load increases. The wear rate is kept low in this state. The other is such that the frictional force and the surface temperature remain constant, even though the load is increased beyond a certain value. In this state, the wear rate increases abruptly with increasing load. The latter phenomenon is discussed in detail and the mechanism of this behavior is proposed on the condition that such behavior appears when the friction force reaches the shearing strength of the material and an apparent area of contact becomes equal to the real contact area. It may be concluded that, pV-limit is regarded as the state where the friction force saturates and remains constant, even though the load is increased (the saturated frictional force state).
Ultrasonic velocities in elastic materials vary with stress applied to them and this phenomenon has been used to determine the third order elastic constants of materials. On the other hand, a new experimental method of stress analysis, named“Acoustoelasticity”, has been developed which utilizes polarized ultrasonic shear waves similar to photoelasticity. This technique, however, has not been put into practice for commercial materials, because of the influences of some uncontrollable factors such as the rate of pre-worked plastic deformation to them. To study a possibility of practical application of this method, the ultrasonic velocities of longitudinal waves propagating perpendicularly to the stress axis of specimens stressed uniaxially were measured precisely in the present work. The materials used were carbon steel“S45C”, aluminum “A3B2”(92%), copper“CuBl-H”(99.9%) and 6-nylon. Furthermore, the modulus deviation calculated from the relation between strains and travelling time deviations was discussed. Using the sing-around method, the ultrasonic velocities were obtained from the travelling time of ultrasonic wave (5MHz) in the 10mm thick specimens. The experimental procedure was as follows: The stress was applied on the specimen and then released abruptly, and the change in travelling time responding to the instantaneous change in stress was measured. The results are summarized as follows: (1) In every material used, the travelling time deviation ΔT/T0 and the modulus deviation ΔM/M0 under a given stress are of the same order of the strain caused by the stress. (2) In copper, the relation between the modulus deviation and the strain is representable by a parabolic curve with little scattering. The results can be expressed by the equation ΔM/M0=Aε-Bε2, where A is a constant relating to the third order elastic constants, and B a constant relating to the density and the pinnig length of dislocation. (3) In annealed aluminum, the modulus deviation measured was in accordance with that expected from the third order elastic constants, but not in steel. The former was explained from that, because of a lot of impurity atoms, the mobility of dislocation would be reduced, and the later seemed to be due to pre-existent structures which would still remain after annealing at a relatively low temperature (250°C). (4) In 6-nylon, the relation between the modulus deviation and the strain shows almost a straigtht line with little scattering.
The electrical conductivity of shock compressed sulfur has been measured at pressures up to 60kb and the mechanism of the electrical conductivity has been discussed. Two kinds of sulfur specimens were used for the experiment. One was made of industrial grade sulfur and the other was of high purity sulfur. A sulfur disk specimen which was sandwiched in between two copper disk electrodes was subjected to shock pressure as the electrode was impacted by a projectile accelerated by a powder gun. It was revealed that both the industrial and the high purity sulfur showed 104 order change in electrical conductivity under the applied pressure, and there was no difference between the two. From the discussion about the conduction mechanism of the shock compressed sulfur, it was clarified that the increase of electrical conductivity with the applied pressure is caused by the increase of electrical carriers activated thermally by shock compression heating, and pressure has only a secondary effect. According to our discussion, the increase in electrical conductivity of sulfur under the applied pressure is not attributable to the mechanism associated with melting as proposed by Berg, but to the one related to the behavior of solid sulfur.
Inhomogeneity and residual stress are two important factors controlling the strength and the fracture in fatigue of induction hardened steels. All the test specimens being used are generally round bars because of a practical use, but they contain many factors which complicate experimental and analytical investigations of the mechanical behavior of these specimens. Therefore, the induction hardened sheets were made in order to obtain basic data for evaluating the strength of those laminated inhomogeneous metals. In this report, as the first step, the inhomogeneity and the residual stress in these sheet specimens were discussed quantlitatively based on the distribution of temperature in the induction-heated and the water-cooled specimens.
A study on the strengthening mechanism for fatigue strength of tufftrided S15CK smooth specimens was carried out. The results obtained are as follows; (1) The fatigue limit of the tufftrided S15CK smooth specimen is about 41kg/mm2, which equals twice the fatigue limit of the untufftrided one. The fatigue limit of the tufftrided specimen can be estimated fairly well from the hardness distribution after fatigued and the residual stress distribution. (2) As the factors which cause the fatigue strength elevation of tufftrided steel, the effects of low temperature quenching, compressive residual stress and supersaturated nitrogen are considered. The ratio of elevation in fatigue limit due to each effect amounts to 10, 15, and 50%, respectively. (3) The compound layer has no effect on the fatigue limit of tufftrided steel.
The low-cycle torsional fatigue tests under multiple repeated strains have been carried out by using an annealed carbon steel S45C, and the test results were used to investigate the applicability of the fatigue fracture criteria based on varying plastic strains as well as on hysteresis energy which was introduced in this paper instead of varying plastic strains. The values of fatigue lives obtained from these fatigue fracture criteria were compared with those calculated from the linear damage criterion. As the main conclusion, the values of cumulative damage obtained from the fatigue fracture criteria based on varying plastic strains and hysteresis energy were slightly closer to unity than those calculated from the linear damage criterion. Therefore, the above two fatigue fracture criteria are considered applicable to the low-cycle torsional fatigue test results under multiple repeated strains. Furthermore, the values of cumulative damage obtained from the above fatigue fracture criteria were almost similar to the test resutls carried out previously by the authors under multiple repeated loads by using a carbon steel S45C.