Flexural strength tests and fracture toughness tests were conducted using 3wt-% MgO partially stabilized zirconia refractories, and a statistical analysis was made on the dependency of strength properties upon porosity. Six kinds of materials with porosity of 5.2, 5.8, 6.6, 7.3, 8.1 and 9.5% were prepared in this study. The flexual strength tests were carried out under 3-point bending, and the fracture toughness of the material was investigated under 3-and 4-point bending using a single-edge notched beam method. The flexural strength and the fracture toughness obtained under 4-point bending decreased in a more porous material, while the variation of porosity did not affect the fracture toughness obtained under 3-point bending. When the Weibull function with two parameters, i.e. shape and scale parameters, was employed to present the flexural strength distribution, the result for the material with higher porosity indicated larger shape parameter and smaller scale parameter. The statistical properties of strength were discussed based on a fractographic observation through a scanning electron microscope. It was found from the observation that the scatter in strength properties of the tested zirconia refractory was dependent on the characteristics of pore distribution, especially the formation of pore clusters, in the tested individual specimens. A fracture mechanics procedure was not necessarily applicable to explain fracture behaviors of porous refractories.
A kinetic model of phase transformation derived statistically from experimental results has been introduced to analyze thermal stresses in a steel plate under the thermo-mechanical control process. The method enables us to predict the stresses more accurately than the conventional method which treats the latent heat of phase transformation as a time-independent function of temperature. By using the method developed in this paper, it was shown that instability of plate flatness in the thermo-mechanical control process is induced by phase transformation and occurs in a heavy plate with thickness about 20mm.
The effects of element partitioning method and temperature model on the precision of one dimensional finite element analyses of transformation and thermal stress were investigated. Partitioning using Gauss integration scheme is found to be effective only in pure thermal stress analysis but not so effective in analysis considering stress induced by transformation and permanent strain. To treat precisely the mechanical property of materials under transformation, each phase generated at each time increment is analyzed part by part individually. The behavior of the phase transformation plastisity is successfully predicted in terms of the plane stress constitutive equation for flat plate.
Extrusion experiments were carried out on mortars to which some gelled substances were added to reduce friction between the materials in the mortars. Both extruding pressure and velocity were measured to elucidate the extrusion characteristics of mortars. The mortars were prepared under several conditions by changing water content, rotational frequency of screw, amount of gelled substance and gelation time. When the amount of gelled substance added was kept constant in the mortars, the extrusion pressure decreased and the extrusion velocity increased with increasing water content. The same behavior was observed when the amount of gelled substance was decreased, contrary to expectation. In the present study, no effect was observed for the extrusion characteristics of mortars when gelled substance was added. It seems that the gelled substance works only as a porous aggregate. Extruding pressure and velocity have been generally known to be prescribed by the extruder operating point which is determined by a cross point of screw characteristics and die characteristics. This concept may be applied to the present mortar-gelled substance system. The die characteristics of screw extrusion were found to be similar to those of piston extrusion.
When an elasto-plastic bar collides with an elastic bar longitudinally, the true stress at its impact end can be detected by using the strain gages glued on the elastic bar with the help of an one -dimensional elastic wave propagation theory. When two elasto-plastic bars collide each other as in the case of the present study, however, the stress at the impact end of the bar can not be detected using the same method, since the plastic deformation occures at the impact ends of the two bars. This paper is concerned with the estimation of the maximum plastic impact stress at the impact end occured by the collision of the two elasto-plastic bars. The integrated value of stress with respect to the whole impact time corresponding to the impulse is introduced to estimate the maximum plastic impact stress at the impact end. The numerical and experimental results are arranged using the integrated value of the impact stress. The results show that the integrated value is maintained constant, when the stress wave propagates in the bar, and increases linearly with increasing the maximum impact stress at the impact end whether the wave is an elastic wave or a plastic one. Accordingly, the integrated value is useful to estimate the maximum plastic stress at the impact end when two elasto -plastic bars are put into longitudinal collision.
The compressive strength of two kinds of titanium alloys was investigated in the wide temperature range from 78K to 723K and the strain rate range from 10-4/sec to 103/sec. It was found that the flow stress increased with the rise of strain rate and with the decrease of temperature, while it was completely independent of temperature and strain rate history. The thermal component of the stress, ≅σ*, was analysed with the thermally activated process concept and was uniquely characterized only with the Larson-Millar parameter ξ*=Tln(ε0/ε) in the relatively low strain rate range. The athermal component of the stress, ≅σμ, was a function of strain only and consequently the whole stress-strain relation at a certain given temperature and strain rate can be predicted. The temperature change due to the adiabatic heat generation was taken into account to evaluate the stress at high strain rates and provided a plausible explanation of the experimental results.
By means of a small punch test technique, the possibility of evaluating embrittlement in the HP-IP steam turbine CrMoV rotors used in service for a long time was investigated. It was found that for the 5% NiCrMoV steel specimens subjected to various aging time, or various degrees of temper embrittlement, the SP energy of the small punch test at -196°C corresponds to the extent of embrittlement, especially the 50% FATT. By adopting this technique to the CrMoV rotors in service, it was recognized that their 50% FATT and Charpy energy at 100°C can be estimated from the SP energy at -196°C. The cause of embrittlement, especially temper embrittlement, can be determined from the observation of fracture surface of the small punch test specimens at -196°C by SEM. The results obtained suggest a possibility to estimate the extent of embrittlement of HP-IP CrMoV rotors in service by a quasi-non-destructive method.
Mechanical properties of GFRP composite materials have been studied using smooth, holed and notched tensile specimens at cryogenic temperatures ranging from 77 to 293K. The maximum tensile fracture stress of the respective specimens increased almost linearly with decreasing test temperature and was related to the cross head speed by an exponential function at each test temperature. The cross head speed sensitivity exponent, m-value, of holed tensile specimens showed relatively higher values as compared with those of the smooth and notched tensile specimens. The notch tensile ratios of the holed and notched tensile specimens were relatively higher by 200 to 300% as compared with those of the estimated values on the basis of the stress concentration factor which was calculated on the ground of elastic theory and the notch shape and size of the tensile specimens, probably due to the stress relaxation at the epoxy resin matrix in the neighborhood of the notched and holed portion of the tensile specimens. The measured maximum fracture stress of the GFRP composite materials was discussed applying the law of mixture and was compared with the calculated values based on the law of mixture. Both the measured and estimated values showed good agreement each other after modifying the previously measured values of the maximum fracture load of the glass fiber strand.
Based on the view that the criterion for the cleavage fracture strength of steels is given by the maximum tensile stress, experiments were performed on a mild steel to study the probabilistic nature of the cleavage fracture stress and fracture toughness. These tests were carried out on smooth round bar tensile specimens, notched round bar tensile specimens and 3-point bend fracture toughness specimens. The cleavage fracture stress in the smooth and notched specimens obeyed the Weibull distribution, but the fracture stress in the notched specimens took higher values than the one in the smooth specimens and the Weibull parameters were different in both specimens. Such effect of notch or stress gradient on the cleavage fracture can be quantitatively explained by statistical FEM analysis based on the Weakest Link Theory. The predicted values of the cleavage fracture stress for the notched specimens which were obtained by the use of the statistical parameters experimentally determined from the tests of smooth specimens, have shown a good agreement with the experimental results. This statistical approach has also been applied to predict the probabilistic distribution of the cleavage fracture toughness. The present experimental results of 60 3-point bend specimens and other reference data on the probabilistic distribution of toughness show the trend to coincide with the theoretical prediction.
Areal inspections by EPMA (electron probe micro-analyser) were made to identify the distributions of impurity elements existing at fracture origins on the ruptured surface of silicon nitride specimens which were tested under tensile load at room temperature. Parallel observations by SEM (scanning electron microscope) were also made to detect the inherent defects at fracture origins. It was found that even if no geometrical defect like a void was observed by SEM at fracture origins, dense segregation of impurity elements such as Fe, Ti and C was found at corresponding sites by EPMA. Therefore, both of the geometrical defects and segregation type defects must equally be treated in the analysis of fracture mechanism. This point was further discussed in this study, and it was concluded that the stress intensity values calculated from the size of both types of defects and the applied load were well correlated with the fracture toughness KIC of this material.
Fatigue crack growth tests were performed on randomly oriented 3% silicon iron in a high resolution, field emission type scanning electron microscope, using a specially designed servo-hydraulic fatigue loading system, and direct, real time observations of growing fatigue crack were made. In the regime with relatively low growth rates, fatigue crack growth behavior was strongly affected by grain orientation, and the growth direction and rate were found to differ depending on grains. However, the effect of grain orientation seems to disappear in the high growth rate regime. It was also found that the grain boundary had an obstructive effect on the transgranular crack growth rate and that the growth rate decreased when the crack tip approached to the grain boundary.
By using S35C/S35C friction welded tubular butt joints fabricated under five different friction welding processes, a series of tests were conducted to determine the effect of welding conditions on joint performance. At first, the hardness measurement revealed that, though the extent of heat affected zone was different from each other depending on the welding condition, no noticeable difference was observed in the maximum hardness value at the weld interface of joint. Then, the results of monotonic tensile tests indicated that the joint efficiencies of four joints were sufficiently high excepting only one joint fabricated under the lowest friction pressure condition. But the above-mentioned results did not necessarily reflect on the fatigue strength of the joints, and their fatigue strength behaviors could be subdivided into two groups as follows: Two types of joints showed lower fatigue strength than that of the base material and appeared to have no fatigue limit in the range of this experiment; and in contrast, the fatigue strength of the other three types of joints well exceeded that of the base material and showed a tendency to have a fatigue limit. Such fatigue strength behaviors seemed to be correlated closely to the magnitude of forging pressure in the friction welding process.
Crystallization behavior was studied in the three series of glasses: (a) (35-x)BaO·xYO1.5·65SiO2, (b) 35BaO·xYO1.5·(65-x)SiO2 and (c) (25+x)BaO·10YO1.5·(65-x)SiO2, with differential thermal analysis (DTA), X-ray diffraction and scanning electron microscopy. Barium silicates with spherulitic structure were main crystals precipitated from these glasses. By comparing DTA curves of powder samples with those of piece samples, it was found that yttrium oxide reduces internal crystallization in the bulk. Stability of glass against heating was evaluated with the interval between the glass transition temperature and the onset temperature for crystallization. Activation energies for crystal growth were determined by the modified Kissinger plot. It was found that YO1.5 increases the stability of glass by increasing the activation energy for crystal growth and this was considered to indicate that YO1.5 strengthens the glass network. In 35BaO·xYO1.5·(65-x)SiO2 series, however, the stability of glass decreased sharply when x exceeded 12mol%, corresponding to the anomalous property change with composition referred to in the previous study.
Two types of epoxy precursors; bisphenol A and phenol novolak, were cured by two kinds of aminophosphazenes; 2, 2-diamino-4, 4, 6, 6-tetraphenoxycyclotriphosphazene (I) and hexa -(4-aminophenoxy)-cyclotriphosphazene (II). Dynamic viscoelastic measurement and thermogravimetric analysis were conducted on these epoxy resins. Compared with the bisphenol A type, the phenol novolak type of epoxy specimens exhibited higher glass transition temperature and higher weight remaining at 300-500°C. The following results were obtained for the epoxies cured by aminophosphazene: E' decreases gradually with the increase of temperature; the value of E" is high and its peak is remarkably broad; and the thermal decomposition initiates at a relatively low temperature compared with the epoxies cured by m-phenylenediamine. These results suggest that the crosslink network is not uniform. The phenol novolak type of epoxy specimen cured by curing agent II exhibits relatively higher E' above 200°C and higher weight remaining in the range of temperature 400-800°C. This type of aminophosphazene is a promising candidate as a curing agent for the heat resistant and flame retardant epoxy resin, although the processability and the uniformity of the crosslink network should be improved.
Waste silica with 5.6μm in the mean particle diameter was mixed with lime, shaped to 7×11×70mm and autoclaved at the range of temperature between 130 and 240°C for several hours under saturated vapor pressure. When the waste silica was mixed with lime in the molar ratio CaO/SiO2=0.4, calcium silicate hydrate gave the greatest bending strength. The observation with a scanning electron microscope (SEM), TG, DTA and X-ray analysis were performed to determine the microstructure, component and crystallinity of the reaction products. The reaction products autoclaved were mainly tobermorite groups and the bending strength reached maximum when the specimens were autoclaved at 180°C for the time more than 48hr. The increase in the ratio of lime tended to raise water absorption and porosity, but to lower bulk density. The thermal expansion of the products was measured by thermomechanical analysis (TMA) in the temperature range of 50 to 700°C. Calcium silicate hydrate expanded slightly from room temperature to 250°C, and contracted at temperatures higher than 250°C, because of its dehydration. The residual silica expanded abruptly at 573°C, because of the transition of α to β type quartz. Calcium silicate hydrate autoclaved at 200°C in the molar ratio of CaO/SiO2=0.4 gave the least thermal expansion and contraction. The addition of sodium hydroxide accelerated the hydrothermal reaction and decreased the water absorption and porosity of the resulting calcium silicate hydrate. The effect of sodium hydroxide on the relationship between the maximum bending strength and autoclave temperature is ambiguous.
Non-linear sin2ψ relations are sometimes observed when the X-ray stress analysis was performed on the materials having preferred orientations. Therefore, the stress determination by the Sin2ψ method is impossible for such case. There are three main factors influencing the oscillation of lattice strain with Sin2ψ as follows: 1) Elastic anisotropy, 2) Plastic anisotropy and 3) Instrumental effect caused by the X-ray optics. This paper deals with the possibility of residual stress measurement for cold-rolled steels having a typical preferred orientation. The results obtained are summarized as follows: 1) The distribution of X-ray elastic constants with respect to Sin2ψ for the (211) diffraction was precisely calculated based on the modified Dölle model and it agreed with the experimental result. 2) The experimental values of X-ray elastic constants also agreed with the analytical values for the isotropic steel by the Reuss model at Sin2ψ=0 and 0.75. Since the X-ray elastic strains at such positions are linear with Sin2ψ, we can determine the elastic stresses in textured steels by using the isotropic X-ray elastic constant by the Reuss model. 3) A remarkable oscillation in the Sin2ψ diagram was observed by the fixed ψ0 method. This oscillation originated from the X-ray optics was analyzed by the numerical simulation. The result indicates that both the X-ray strains at Sin2ψ=0 and 0.75 on the (211) diffraction agree with those by the fixed ψ method. Thus, we proved that the residual stresses in textured steels can be measured by using both X-ray strains at such values of Sin2ψ.