The biaxial creep deformation behavior of celluloid softened by heating was tested under various stress patterns which consisted of loading, unloading and stress reversal processes. The results obtained in the experiments are summarized as follows. (1) Just after the abrupt change of stress pattern, the stress vector shifted its direction from that of creep strain increment vector. Their directions became coincided with each other with the lapse of time. (2) The angle between these two vectors became large during the partial unloading process. (3) As the stress level became high, the length of the creep strain curve up to the point where the directions of both vectors coincides became long. (4) Under the stress reversal, the variation of the corresponding creep strain component became especially large. (5) For the stress change with a mirror image pattern, the isotropic characteristics of creep deformation appeared as the number of stages increased.
The crack growth initiation behavior in delayed failure under superposed static or dynamic mode II stress on mode I static stress was investigated on the precracked specimens of Ni-Cr-Mo steel that was quenched and tempered at 673K. When the superposed mode II static stress intensity factor KII or the range of mode II dynamic stress intensity factor ΔKII on the static stress intensity factor KI was relatively small, the crack initiation time ti was longer than that when KII or ΔKII was zero. When KII or ΔKII was large, however, ti decreased with an increase in KII or ΔKII. The crack propagation direction under superposed KII or ΔKII on KI is dependent on fracture mechanisms. When hydrogen embrittlement governed the mechanism, the crack propagation angle θp once increased with an increase in ΔKII /KI and decreased with a further increase in ΔKII/KI. In the case of fatigue θp was small when ΔKII/KI was small and increased with an increase in ΔKII/KI. The crack propagation under superposed ΔKII on KI in water occurred by the hydrogen embrittlement type when ΔKII/KI was small, and changed to the fatigue type when ΔKII/KI became large.
A three dimensional boundary element method was developed for predicting electro-galvanic field responses due to anodic-cathodic interaction. Although the governing equation is linear (Laplace), the boundary condition, which is enforced based on experimentally determined polarization curves, is generally non-linear. Therefore, Newton-Raphson iterative procedure was employed in the present prediction method. An experiment was also conducted on galvanic corrosion of a cylindrical vessel which was made of cast iron and stainless steel, and filled with 0.0165% NaCl solution. In addition, the effect of cathodic protection by placing a small piece of aluminum plate on the cylindrical vessel was examined experimentally. The validity and usefulness of the present prediction method were demonstrated by comparing the calculated results and the experimental data.
In the previous study, the influence of cyclic frequency on the fatigue crack growth rate in liquid environments was investigated. The corrosion fatigue crack growth at high frequencies was found to decrease to almost the same value as that in the laboratory air, possibly due to the gaseous phase produced by the cavitation bubbles gushing at the tip of the crack. In this study, the influence of cavitation bubbles on the fatigue crack growth behavior in synthetic seawater was investigated in terms of stress waveform. When the stress waveform with a constant period changed its rising rate to a high value, the cavitation bubbles were observed at the tip of the crack, and the corrosion fatigue crack growth rate approached to the value obtained in laboratory air.
Bending fatigue tests were carried out on the specimens coated with VC by immersing in a molten borax bath and on the base specimens hardened by quenching and tempering. The residual stresses in the carbide layer and substrate, the hardness and grain size of the substrate were also examined. In the VC layer, the compressive residual stress of 1000-2000MPa appeared, while in the substrate of the VC coated specimen and the surface of the hardened base specimen, the tensile residual stress of 0-400MPa appeared. This tensile residual stress was generally higher in the VC coated specimen than in the hardened base specimen. The bending fatigue limit of the VC coated specimen was influenced not only by the hardness but also by the residual stress in the substrate. In the case that the hardness is constant, the bending fatigue limit is linearly correlated to the residual stress in the substrate, and also this correlation holds good between the bending fatigue limit and the residual stress on the surface of the hardened base specimens. The bending fatigue limit was not increased by tempering at high temperature which lowers both the tensile residual stress and the substrate hardness. On the other hand, diffusion soaking was effective in increasing the bending fatigue limit, because the tensile residual stress becomes lower and the substrate hardness becomes higher.
Fatigue crack propagation characteristics at the stage 2 of spheroidal graphite cast iron and high strength steel with equivalent tensile strength and yield stress have been studied and compared by means of fracture mechanics and fractography. In the high ΔK region, the stage 2c cracks consisting of cleavage fractures were found only in the spheroidal graphite cast iron. Thus, its fatigue propagation rate was faster than that of high strength steel without the stage 2c cracks. In the medium ΔK region, on the other hand, both materials showed the fatigue crack propagation at the stage 2b, although the propagation rate of the spheroidal graphite cast iron was faster than that of the high strength steel due to the smaller Young's modulus of the former. Nevertheless, the propagation rates of both materials were the same, when their Young's moduli and crack closures were taken into consideration. In the low ΔK region, both materials showed the fatigue crack propagation at Stage 2a, although the propagation rate of the spheroidal graphite cast iron was slower than that of high strength steel, different from the high and medium ΔK regions. This is due to a more significant crack closure caused by the roughness of fracture surface and the formation of oxide in the cast iron than the steel. Moreover, the crack branching phenomenon was found in the cast iron where Kmax=16 to 33MPa√m. In the case that this branching phenomenon occurred in the low ΔK region, the propagation resistance of fatigue cracks became higher.
The rolling contact fatigue tests were carried out on a high manganese austenitic cast steel to investigate the influence of grain size on the rolling contact fatigue strength by using a Nishihara-type wear testing machine under a lubricated state with a relative slip ratio (-9%). The main results obtained are summarized as follows: (1) The rolling contact fatigue limit of the high manganese austenitic cast steel increased remarkably with reducing grain size. (2) The effect of heat treatment on the reduction of grain size was almost equal to that of forging. (3) The rolling contact fatigue limit of the high manganese austenitic cast steel increased remarkably with work handening generated by repeated contact stress. (4) A crack in the high manganese austenitic cast steel propagated through the large cyclic-strained zone in the parallel direction with the contact surface. Its depth is in the vicinity of the region (Z0) having the maximum stress value calculated by the maximum shearing stress theory.
The effect of stress frequency on fatigue crack initiation and propagation in SM50A steel, which has strain rate dependence in the plastic region, was studied. Fatigue crack initiation and propagation tests were carried out at 0.02Hz, 0.2Hz, 1Hz and 20Hz, at R=0, at room temperature, and the number of cycles to crack initiation and the crack propagation rate were measured. The local stress-strain behavior at the part of stress concentration was analyzed and the simulation of fatigue crack propagation was performed by FEM employing the elasto/visco-plastic overlay model. The comparison between the experimental results and the visco-plastic strain behavior calculated by analysis was made. The following results were obtained. (1) The number of cycles to crack initiation and the crack propagation rate depend on the stress frequency in such a way that for lower frequency, a shorter life of the crack initiation and a higher crack propagation rate were obtained. (2) The dependences of the crack initiation life and the crack propagation rate on stress frequency can be explained by the visco-plastic strain range based on the material viscosity in the plastic region at the notch root and the crack tip. (3) A parameter closely related to fatigue crack initiation and propagation rate is the visco-plastic strain range at the notch root and the crack tip in visco-plastic materials.
In order to clarify the effect of compressive stress on the fatigue crack propagation in polycarbonate, fatigue crack propagation tests with a negative stress ratio and with a single compressive overload were carried out on the pre-cracked specimens. The results obtained in this experiment are summarized as follows: (1) The compressive stress range in the cyclic loading with a negative stress ratio causes the acceleration in fatigue crack propagation under constant-amplitude loading. (2) This phenomenon can be explained mainly on the basis of the crack closure concept, but when the negative stress ratio is very high, other factors may also contribute to this phenomenon. (3) The accelerated crack extension during a tensile overload is not affected appreciably by the negative stress ratio cycling or a single compressive overloading preceding the tensile overload. (4) The retardation effect due to the tensile overload is not reduced by subsequent negative stress ratio cycling or a single compressive overloading. This fact suggests that the retardation in polycarbonate cannot be attributed to the plasticity-induced crack closure.
The residual stress generated during the molding process of thermosetting resins gives a serious influence upon the mechanical properties. This residual stress is generally classified into two groups: one produced by shrinkage in the curing reaction of monomers, and the other produced by non-uniformity of the temperature distribution in the cooling process. This paper is concerned with the influence of residual stress by rapid cooling on flexural fracture behavior of epoxy resin. It is explained by using thermoviscoelastic model and confirmed experimentally that the large frozen strain as well as residual stress are generated by the thermoviscoelastic behavior of epoxy resin when it is cooled rapidly. The residual stress have a large effect on the flexural fracture behavior of epoxy resin, which is independent of the frozen strain.
In order to investigate macroscopically the effects of crystallinity and temperature on the fatigue crack growth behavior of polypropylene resins, the load controlled fatigue test was conducted and the macroscopic consideration of the fatigue crack growth behavior was made in this paper. The results obtained are as follows: (1) The fatigue strength of the notched specimen at a given temperature increased with an increase in crystallinity and mean size of spherulite, but it was influenced by the crystallinity more than the mean size of spherulite. The fatigue strength also increased with a decrease in temperature. The slope of S-N curves was gentle at temperatures near the glass transition point. (2) In the fatigue crack growth process, two kinds of damage zone occurred at the crack tip, and the size of the damage zone in the load direction decreased with a decrease in temperature. The mechanism of fatigue crack growth was made clear for the specimens with different crystallinities at room temperature. (3) The fatigue crack growth curves were divided into three types. The type I is a smooth curve at room temperature and the type II is a stepwise curve at temperatures below 20°C. It is the type III that the unstable fracture occurs suddenly when the stable crack grows a little and the crack growth curve is almost flat at lower temperatures.
Prior to the launching, several certification tests are carried out for satellite and rocket on-board components. The ground test requirements are specified strictly to verify their reliabilities and abilities to mission objectives. But it is not easy to define ground test parameters such as load patterns, levels and durations in the ealier stages of whole system development. These parameters would be different in general from those expected in the preliminary design stage. This work is concerned with the effect of various load patterns on the fatigue strength of mission equipments. A method of evaluationg fatigue damage by using the fatigue damage accumulation (FDA) ratio is proposed first, then the FDA ratio was computed for different types of load patterns. Random load time history which has a given power spectrum density function was generated by digital computer simulation. Two dimensional cycle counting by an advanced type of the range-pair-mean method was adapted. Miner's rule was applied to calculate cumulative fatigue damages. This work concentrates on the FDA characteristics of various random load patterns which have the same power level. Simulation results showed that the load patterns with predominant frequency have about the same fatigue damage accumulation (FDA) ratio as the pattern without predominant frequency if both of them have the same frequency band width. The relationship between the irregularity factor and FDA ratio was also discussed.
Granitic rocks have an orthotropic (rhmbic) elasticity. In quarryman's terminology, these anisotropic planes are called a rift plane, a grain plane and a hardway plane in the order of ease of splitting. The axes R, G and H are normal to these planes. In the anisotropic body, three independent elastic wave velocities in all directions exist. These are the velocities of the quasi-longitudinal wave involving shear motion, and of two quasi-shear waves involving dilatational motion. These velocities can be determined by nine stiffness constants for orthtropic materials with Kelvin-Christoffel's equation. In this study, two octadecahedrons and three trioctagonal prisms whose axes coincide with H, G and R of Oshima granite were used to determine the velocities of various directions of propagation and polarization. The nine independent stiffness constants were determined by the nine independent velocities. The polyhedron was loaded up to 200MPa under hydrostatic pressure. All components of the stiffness constants increased with an increase of pressure, showing that the anisotropy of the granite should be attributed to the pre-existing microcracks.
Alkari-silica reactions have a harmful effect on some andesites. In the present paper, the alkali-silica reactions in andesitic rocks were examined by means of optical and X-ray diffraction analyses and the standard test with mortar bars (ASTM C227). Furthermore, by the microscopic examination of the dyed thin section made from a mortar bar, various types of reaction products including cracks and spherules were identified and distinguished. The contents of alkari-reactive minerals such as cristobalite, tridymite, and glass, which are present in andesites, are irregular and change very much with the location. The results of the test on mortar bars had no direct correlation with any kind of alkali-reactive minerals. It is difficult to determine the role of these alkali-reactive minerals in the expansive alkali-silica reaction, because there are many factors in the reaction.
The mortar-bar expansion test (ASTM C227), chemical test (ASTM C289) and measurement of specific gravity were performed on the aggregates of which the mineral characteristics were discussed in the previous report. The cement used for the mortar bar test was an ordinary portland cement with an alkali content of 1.6% (equivalent Na2O) adjusted by NaCl reagent. The results of the tests on 27 mortar bars showed that the rate of expansion was in the range of 0.02-0.65% in 6 months. Twenty-one aggregates were classified as being reactive according to the ASTM criteria. In the chemical test, the determined concentration of silica in solution ranged from 190mmol/l to 670mmol/l. The reduction in alkalinity of the NaOH solution ranged from 65mmol/l to 290mmol/l. All of the 36 aggregates belonged to the reactive group. The specific gravity of the aggregates was in the range of 2.19-2.68. Although no relationship was observed between mortar bar expansion and the result of the chemical test, a negative correlation was found between mortar bar expansion and specific gravity. To prevent excessive expansion, the isopleth map of the rate of expansion was drawn based on the test results on a number of mortar bars to cover variables in alkali content of the cements and in mixing ratio of the reactive aggregate of andesite and the nonreactive aggregate of greywacke.
Ferrous (mild steel and stainless steel) and non-ferrous (Ti and Ni) alloy plates were diffusion-welded under thermal cycles including the solid-state transformation temperatures of the mild steel and titanium by placing them in the cylindrical electrodes, and the relation between the shear strength of joints τj and the ultrasonic echo intensity from the joint interface was investigated. The relation between the formation of reaction layer at the joint interface and the ultrasonic echo intensity was also investigated. In the mild steel/Ti and stainless steel/Ti joints, a brittle reaction layer was formed at the joint interface, and no relation between τj and the area of joints Aj (the area of good welding estimated from the fracture surface) was found. No relation between τj and the diameter Du estimated from the distance of low ultrasonic echo intensity from the joint interface was also found. In the mild steel/Ni joints, no reaction layer was formed at the interface. The strength of the joints varied linearly with the area Aj, and had a quadratic relation with Du. When no reaction layer was formed at the interface, the echo intensity ratio (Re) was in good agreement with the reflectivity (r) at the interface over the joined region. When the reaction layer was formed at the interface, the variation of Re disagreed with that of the calculated echo intensity ratio.
The theory of acoustoelastic effects in soft-ferromagnets is formulated. The materials are assumed to be isotropic and magnetostrictive. The initial magnetization curve is set to be linear with respect to the applied magnetic field. General acoustoelastic formulae are obtained taking account of the effects of magnetic field, magnetostriction, and applied stress. The stress application has direct acoustoelastic effects and indirect effects due to the stress dependence of magnetic susceptibility. Special cases of simple geometries giving rise to analytical solutions for the elastic wave speed are examined. For the case of uniaxial stresses, magnetostriction causes a shift of stress vs. acoustic birefringence curve, and a slight nonlinearity of the curve arises from the stress dependence of the susceptibility. The case of plane stresses is also studied to show a possibility to specify both components of the plane stress using elastic waves propagating normally to the plane of a plate.
The present paper describes the electromagnetic impact tests of such cryogenic materials as a Ti-6Al-4V alloy and a 25Mn-5Cr-1Ni steel. The tests were performed in a steady magnetic field of 6 T at liquid helium temperature (4.2K) by using a newly developed machine. The new impact tester, which is composed of a superconducting magnet, an electric pulse generator and a measuring system, can apply electromagnetic in-plane bending moments up to approximately 1MN·m to a single-edge-cracked specimen. The deformation behavior of the specimen was monitored by a newly designed magnetic pick-up-coil and cryogenic strain gages of the foil type. The transient responses of the deformation rate and strains were obtained by these devices. The critical peak value of the applied current pulse, or the intensity of the applied electromagnetic force, which causes the fracture or the crack growth, decreased with an increase in fatigue precrack length for the relative precrack length of a/W<0.5. The critical peak current approached to a constant value for a/W>0.5 for each material. The value of the critical peak current was 3-3.6 times higher for the high Mn steel than that for the Ti alloy. This result implies that the high Mn steel has higher resistance to the electromagnetic impact force than the Ti alloy.