The piston crown of a diesel engine is subjected to both thermal and mechanical stresses. The thermal stress arises from temperature gradients in the piston crown, and the mechanical stress to gas pressure. These stresses are superposed each other. The fatigues under superposed wave stresses have been studied considerably in the past and several methods for fatigue life prediction under such stresses have been proposed. In this paper, the fatigue behaviors under superposed wave stresses are investigated and a new method to predict the fatigue life is proposed. An example of the prediction on the piston crown of a diesel engine is also presented.
For the preparation of creep data sheets, a systematic and coordinated test program was planned at National Research Institute for Metals in 1964. The plan was to obtain creep and rupture strengths up to 100000hrs by performing long-term tests on high-temperature materials produced in Japan. To carry out this program a new creep laboratory was built at NRIM in 1967 and the installation of testing facilities was completed in 1968. According to the program the tests were started in 1968 and the program is now under way for about 30 kinds of steels. More recently, on 6 kinds of steels (1Cr-1/2Mo, 11/4Cr-1/2Mo, 21/4Cr-1Mo, 18Cr-8Ni, 18Cr-8Ni-Ti, and 18Cr-12Ni-Mo steels), the data up to 10000hr rupture-times were obtained, and these data were published as NRIM Creep Data Sheet No. 1∼6. The creep rupture-times of 21/4Cr-1Mo, 18Cr-8Ni and 18Cr-8Ni-Ti steels showed larger scatter than those of the other 3 kinds of steels. The elongation of ruptured specimens decreased with increasing rupture-time and showed relatively large scatter for 18Cr-12Ni-Mo steel. Furthermore, with regard to the 100000hr rupture stress a comparison was made between the estimated values obtained from the regression curves for logarithmic stress versus logarithmic rupture-time plots and other published data.
In this report shown are the test results of short-time bursting of thick-walled tubes of a heat-resisting alloy, 15-15N, subjected to internal pressure at high temperatures up to 700°C. The diameter ratio of the test tubes are from 1.75 to 7. The experimental bursting pressures are compared with the estimated pressures by several well-known design formulae. The appearances of burst tubes are also examined. The results obtained are as follows: (1) Among the formulae proposed for room temperature, the formulae by Manning and by MacGregor et al. are not quite applicable to the high temperature range in spite of tedious calculations. (2) The experimental values yields a linear relationship in the ultimate pressure-ln (diameter ratio) plots, and Faupel's formula gives a good agreement with the experimental results. (3) The crack initiates at the inside tube wall and propagates toward the outer part to cause bursting of tube. The appearance of burst surface at the cross-section indicates cleavage or shear fracture for the tubes of small diameter ratio. For the tubes of larger diameter ratio the cleavage fracture produced at the inside surface transforms into the shear type which propagates along the outward logarithmic spiral direction from the top of a large cleavage crack.
In the last few years considerable attentions have been given by many investigators to the couple-stress theory. The couple-stress theory considers that, within an elastic body, the surface of each material element is subjected not only to the normal and tangential forces but also to the moment per unit area. Such an assumption seems appropriate for materials with granular or crystalline structure. Applying Mindlin's and Nowacki's theories, in this paper, we solved the problem about the effect of couple-stress on the thermal stress distribution in a polygonal prism with a central circular hole under a stationary temperature distribution. The result shows that, by taking couple-stresses into account, the stress-concentration factor is obviously reduced from the value acceped in classical theory.
The finite element method was applied to formulate the elastic-plastic analysis of quenching involving the transformation by taking considerations of the dependence of thermal expansion coefficient on both the temperature and the cooling rate. The stress-strain incremental relation including the term due to quenching was deduced and then used to obtain the equilibrium equation capable of representing the effect of the transformation. As an example of the formulated theory, the numerical calculation was carried out for 0.43% C carbon steel circular bar with 60mm diameter during water-quenching. The calculated residual stresses agreed well with the experimental values measured by Sachs' boring-out technique.
A method to estimate the deformation due to thermal ratcheting under multiaxial stress was presented from the viewpoint of thermoplasticity by employing the general yield function associated with the kinematic hardening rule. The analysis developed here indicates that the ratchet strain of a sufficiently work-hardened material under varying stress and temperature can be predicted from the data of the progressive strain under uniaxial cyclic stress at constant temperatures. Also the comparison was made with the experimental results of low carbon steel subjected to uniaxial cyclic thermal stress combined with mechanical steady stress.
The strain controlled thermal fatigue test was carried out on HK-40 and 25Cr-35Ni centrifugal casting tube materials being used for heat transfer tubes of chemical apparatus at about 1000°C. In addition to the general thermal fatigue properties, the evaluation of creep damage, behaviour of welding joint and metallurgical crack appearance were also investigated. The results are summarized as follows: (1) The results of thermal fatigue properties fall on straight lines when plotted ΔεT against Nf in the logarithmic scale. The slope of the line for Tmax=1100°C differs from those below Tmax=1050°C. This difference is considered to arise from the oxidation effect. (2) It has been generally said that as the damage of trapezoidal wave pattern rupture occurs when Φh(=Φf+Φc) reaches unity. However, the present results show that it occurs when Φh is more than unity at low temperatures and less than unity at high temperatures. Furthermore, the considerable effects of creep recovery and oxidation besides creep damage and fatigue damage are noticed.
Many criteria for thermal fatigue failure have been proposed, but almost all of them seem to be little authorized. Both the combined effects of temperature and strain as well as the wide variations of test methods make it difficult to find a general rule. In order to evaluate the effect of temperature on the relation between the plastic strain range and the number of cycles to failure, the authors carried out thermal fatigue tests with a low carbon steel at three different mean temperature levels and temperature ranges. The experimental results were compared with the predicted life of the materials obtained from the low cycle fatigue data at constant elevated temperatures. The test system contains a newly-devised extensometer and a servo-control equipment. It can change the strain ranges independently from the temperature ranges, while keeping the proportionality between instantaneous strain and temperature constant. The results obtained are summarized as follows: (1) The relation between the plastic strain range Δεp and the number of cycles to failure Nf under fixed Tmax and Tmin can be described by Δεp·Nfβ=const., where β is a constant much greater than 0.5. (2) Both the increase of mean temperature level and the increase of the temperature range give the same effect on the decreasing fatigue life of low carbon steel under the present test conditions. (3) The life predicted from low cycle fatigue at elevated temperatures under the assumption of linear damage theory coincides with the experimental values at relatively low temperatures, but it is more than the experimental values at temperatures higher than 450°C in the case of low carbon steel.
Fatigue under varying temperature conditions and thermal fatigue are important problems for designing machines and plants used at elevated temperatures. In order to deduce a fatigue life rule under these complex conditions, it is necessary to find the physical changes which correspond to the“damage”, and then to clarify the effect of applied conditions on these changes. The authors examined the feature of cracks in the partially fatigued specimens and the effect of two-step temperature level on fatigue life. As the result of this investigation, the following conclusions have been obtained: (1) The number of cycles to failure of a low carbon steel decreases monotonously with increasing temperature level from R. T. to 600°C for a total strain range of 1%. (2) Submacro-cracks (30∼50μm depth) initiate and grow in the early stage of fatigue test at the temperature level of 500°C, but at R. T. level they do not appear until the later stage. (3) The order of application of the two temperature levels affects the fatigue life only when the upper temperature exceeds 400°C. (4) The total crack area seems to be a more realistic measure of damage, and it is useful for estimating fatigue life under the two-step temperature level test.
The low-cycle fatigue tests were conducted at elevated-temperatures on two alloys, Type SUS 32HP (equivalent to AISI 316) stainless steel and 1Cr-1/4 Mo steel, SCM3, and the effect of holding time at the absolutely maximum tensile or compressive strain was investigated. A considerable reduction in their fatigue lives was observed when the test temperature was raised and the holding time was increased. The linear fraction rule between creep and fatigue damage was compared with the experimental results, and it was found that there lie several serious problems in this rule, especially when applied to cyclic softening materials such as SCM3 steel. The Endo's theory on corrosion-fatigue was modified to develop the following life-evaluating equation, which takes consideration for the relationships among the cycle-dependent effect by fatigue damage, the time-dependent effect by oxidation, and their combined effect. (1+B·t0n1·Nfn1)·εpR·Nfα0=C0 where, t0, εpR, and Nf are the holding time during a half cycle, the plastic strain range, and the number of cycles to failure, respectively, and B, n1, n2, α0, and C0 are constants. This equation was shown to represent the experimental results well. The time-dependent effect may be interpreted to include other supplementary effects such as creep damage and deterioration of materials.
In order to collect experimental data of the fatigue strength of materials, the rotating bending fatigue tests were carried out on two kinds of austenitic stainless steels (Type 304 and 316) at room temperature, 400, 500, 600 and 700°C with the frequency of 7500r.p.m., and their fatigue strength at 108 cycles was obtained. The conclusions obtained are as follows: (1) Endurance limits were clearly found in the temperature range from room temperature to 600°C. At 700°C a clear endurance limit seemed to appear below 107 cycles, but the fatigue fracture started to occur after 107 cycles. (2) The fatigue strength of SUS 316B at 108 cycles is larger than that of SUS 304B at each test temperature. (3) The ratio of fatigue strength to 0.2% proof stress is 1.2∼1.7 at 400-600°C. This ratio is larger for SUS 316B than for SUS 304B. (4) The coaxing effect was distinctly observed in the temperature range from 400 to 600°C. At 700°C, it was somewhat observed during the test after 9×105 cycles but it disappeared after 108 cycles. (5) The above phenomena commonly observed for the two steels are considered to arise from the same origins. They are the strengthening of the material caused by aging during the fatigue test at 400-600°C, and the softening due to overaging after 107 cycles at 700°C.
The effects of changes in shape and size of specimens on creep and rupture properties of low carbon steel were investigated on the plate specimens having various thickness (t0), breadth (b0) and parallel length (L0). t0 ranged from 2 to 30mm, b0 from 30 to 60mm, and L0 from 32 to 180mm. L0/√A (A: area of cross section) varied between 2 and 13.7. Creep and rupture tests were carried out at 450°C and 550°C. It is shown that, with increasing L0/√A, the rupture time decreases in the case of plate specimens when L0/√A≤6, but it is constant when L0/√A>6. On the other hand, it increases with L0/√A in the case of round specimens when L0/√A≤6. With increasing L0/√A, the minimum creep rate increases in the case of plate specimens when L0/√A is small, but it is constant when L0/√A>6. The changes in t0, b0 and L0 of the plate specimens affect both the rupture time and the minimum creep rate little if L0/√A is maintained constant. In the case of short rupture time, the plate specimens with the dimensions of b0/t0>10 show the fracture oblique to the tensile axis across the broad sides. The angle of inclination made by the line of fracture is approximately 60°.
The creep rupture strength and metallurgical microstructures of tube and pipe steels used for high-temperature and high-pressure service were investigated. A total of 127 kinds of test materials were selected from 17 groups of carbon steel, low alloy steel or austenitic stainless steel. The test specimens were taken from the wall of the tubes and pipes which were commercially produced in our tubing plant. The average creep rupture strength of each steel was obtained from the creep rupture test performed for the period longer than 10000hrs with a multiple-type testing machine. The metallurgical microstructures after the creep rupture test were investigated by optical- and electron-microscopy as well as X-ray and chemical analyses of the extracted residues. The results obtained were used to discuss the relations between creep rupture strength and microstructures. In carbon steel, active N and Mn are effective for strengthening but C and pearlites of both lamellar and spheroidal types are not so effective at high temperatures in an extended time. In low alloy steel, the carbide distributions and carbide types play important roles in the strengthening. MC carbide in austenite stainless steel increases the creep rupture strength but at high temperatures this effect disappears.
From the results of the tensile creep rupture test on notched plate specimens of low carbon steel at 450°C, the following conclusions have been derived. (1) The level of equivalent stress at the notch bottom of the notched specimen under a fixed nominal stress is lower than that on the cross section of necked part of the unnotched specimen during the greater part of rupture life. The equivalent creep rate at the notch bottom of the notched specimen under a fixed nominal stress is also smaller than that of the unnotched specimen. Therefore, it is considered that the time for crack initiation in the notched specimen under a fixed nominal stress is longer than that in the unnotched specimen. (2) Both the level of hydrostatic stress at the notch bottom of the notched specimen under a fixed nominal stress and the rate of its increase with lapse of creep time are lower than those on the cross section of necked part of the unnotched specimen during the accelerating creep stage. Therefore, it is conjectured that the crack propagation rate of the notched specimen under a fixed nominal stress is lower than that of the unnotched specimen. (3) Rupture time versus minimum creep rate curve at the notch bottom of the notched specimen differs from that of the unnotched specimen. On the assumption that the crack initiation time versus minimum creep rate curve at the notch bottom of the notched specimen coincides with that of the unnotched specimen and creep rupture of the material occurs when the total length of cracks reaches to a limited length, it is considered that the time for crack initiation in the notched specimen is longer than that in the unnotched specimen. It is also conjectured that the crack propagation rate of the notched specimen is smaller than that of the unnotched specimen.
The creep rupture tests on the precracked specimens made of low carbon steel and OFHC copper were carried out at 450° and 200°C, respectively, with microscopic observations of the crack propagation process. The precracked specimens of low carbon steel showed a remarkable “notch strengthening”, the notch strength ratio of which was as much as 1.5 over the whole range of life investigated, while the OFHC copper specimens showed the notch strength ratio of only 1.1 to 1.2. For low carbon steel, cracks started to grow at about 60% of the expected rupture life and propagated in a mixed trans- and inter-granular mode. Large crack opening was observed. On the other hand, for OFHC copper the initiation of crack growth was observed as early as at about 40% of the expected rupture life. The cracks propagated in an intergranular mode with less crack opening. This difference of the crack propagation behaviors is thought to be responsible for the remarkable difference in the notch strength ratio between low carbon steel and OFHC copper.
A study of the mechanics of creep fracture of metals and alloys was carried out on the basis of the following assumptions. (1) The nucleation rate of a microcrack and the rate of creep crack propagation are proportional to the rate of a locally concentrated microstrain. (2) The rate of a local microstrain can be expressed by an equation as a function of macroscopic stress, strain and temperature. (3) The locality and the inhomogeneity of such microstrains come from easy concentration of creep deformation along grain boundaries and/or sub-boundaries, in the vicinity of precipitated particles and so on. As the result, the propagation rate of a grain boundary crack was given analytically by a power function of a net section stress. This was substantiated by the experiment on the creep crack propagation in notched specimens of 1Cr-1Mo-1/4V steel at the test temperature of 600°C.
An attempt is made to explain the growth process of creep cavities wherein the diffusion mechanism is not applicable. When a screw dislocation reaches grain boundary, the dislocation starts to move on the different slip plane by cross-slip. But such movement of the dislocation is interrupted at a precipitated particle located near the grain boundary, and the dislocation moves back to the grain boundary by another cross-slip. Thus the dislocation moves along the grain boundary with successive cross-slip, until it is absorbed by a cavity in the boundary, resulting in growth of the cavity. The rate of the growth of cavity is calculated and the relationship between creep rupture time and creep rate is determined. The effect of coarsening of precipitate particles upon the creep rupture time is discussed. In the case of no particle coarsening dung creep, the creep rupture time tr is inversely proportional to the minimum creep rate εm. However, tr is proportional to εm-0.6 when the coarsening of particles takes place during creep. The theory is well consistent with the experimental data obtained for type 316 and 347 stainless steels.