Fatigue tests were conducted at a frequency of 20Hz or 5Hz under push-pull or reversed torsion or combination of these two loading modes. Temperature of specimen rose by heat generation during cyclic stressing at 20Hz but was kept at room temperature by air-cooling throughout the testing at 5Hz. The variations of fatigue strength and strain range resulted from heat generation were studied in each loading mode. The heat generation became greater in the order, torsional, combined, push-pull loading, which resulted in shortening of the fatigue life at higher stress amplitude. The fatigue strength under combined loading was able to be evaluated by using S-N diagrams derived on the basis of Gough's ellipse quadrant theory except that the slope in the finite fatigue life region varied depending on temperature rise. Mises' equivalent plastic strain range (Δεp)eq increased rapidly at an early stage of fatigue life and decreased during subsequent stress cycles. In the case of heat generation, the change of (Δεp)eq was amplified and its average in fatigue process was lowered. The relationship between the maximum of (Δεp)eq and fatigue life was expressed by a straight line with a narrow scatter band on a log-log diagram, irrespective of the heat generation.
A study was made on the effect of hardness on fatigue properties of high strength steels. Rotating bending fatigue tests were carried out on four different carbon steels tempered at four temperatures. The results are summarized as follows; (1) Origin and location of fatigue fracture could be classified by hardness. (2) Even for high hardness steel showing fish eye fracture, fatigue limit was proportional to hardness under the same tempering condition. (3) Fatigue limit could be estimated as a function of hardness and tempering temperature.
The fatigue fracture surface of high strength steels often shows a typical mark like fish-eye. A nonmetallic inclusion is commonly found at the center of fish-eye mark. This implies that the nonmetallic inclusion at the center of fish-eye of a specimen fractured by tension-compression fatigue test is the maximum one existing in the specimen. It was confirmed, from the experiments on many sample of a high-speed steel, that the size (the square root of projected area: √area) of the nonmetallic inclusions at fish-eye followed the statistics of extreme values. Moreover, it was found that the fatigue fracture stress at these inclusions could be predicted by Murakami et al's equation which was proposed on the basis of fatigue data on small artificial defects. The lower limit of a large scatter band of the fatigue strength was also predicted by combining the statistical distribution of extreme values of √area and Murakami et al's equation. It was shown that the prediction was in good agreement with the experimental results.
Spheroidal graphite (SG) cast iron is a typical porous material which contains numerous small defects (graphites) in the matrix. In order to clarify the geometrical parameter for the quantitative evaluation of the effect of surface graphite on the fatigue strength of SG cast iron, rotating bending fatigue tests were carried out on the plain- and notched-specimens of as-cast FCD60 and FCD70. It is concluded from the microscopic observation that the fatigue limit is controlled by the maximum size of surface graphite defined by the square root of the projected area of graphite (√area*max). The value of √area*max in a definite surface area of the specimen can be predicted by the statistics of extreme values. The predicted values of √area*max approximately agree with the critical size of artificial notch which does not act to reduce the fatigue strength. The value √area*max is in general considered to be the controlling geometrical parameter for the prediction of fatigue strength of porous materials.
In surface treatments, compressive residual stress developed near the surface layer is one of the important factors to improve the fatigue crack propagation life as well as the crack initiation life. In this study, the effect of residual stresses induced by case-hardening on the crack propagation life was analyzed. A crack was assumed to initiate from a surface or a subsurface defect, and the location of the defect to give the shortest life was determined. When a crack propagates under the compressive residual stress field, a partial crack surface contact occurs. In the present life evaluation, this effect was considered. Although the residual stresses are effective to improve the crack propagation life for a surface defect, they bring an undesirable effect for a subsurface defect. The analysis shows an optimum residual stress distribution to give the most desirable effect on the propagation life for both defects.
In the previous paper, the fatigue strength data of notched specimens of steels had been analyzed by distinguishing the fatigue limit for crack initiation, σw1, and that for fracture, σw2, with the aid of Nishitani's critical notch radius ρ0, at the branch point, i.e., at the critical point where a non-propagating crack appears. In this study, S-N (P-N, S-T) data of rotary bending tests of sharp (notch radius ρ_??_ρ0), deep (notch depth t>>√ρt/5), notched specimens of steels were extracted from JSMS Data Book and analyzed by fitting a bi-linear type S-N curve comprising of inclined and horizontal straight lines on a Log-Log diagram, using Nishijima's method. By taking the stress intensity factor K of the circumferential crack with notch depth t instead of the stress σ in S-N curves, and by using the following correction factor κ for low strength steels, i.e. κ=σB/σC for tensile strength σB<σC=940MPa, κ=1 for σB≥σC, S-N Data of sharp, deep notched specimens of both low and high stength steels could be fitted by a bi-linear type curve on a Log (K/κ)-Log N diagram (Fig. 4(e)), with the coefficient of variation 14.0%, while the coefficient of variation for the Log (σ/σB)-Log N diagram (Fig. 3(c)) of corresponding smooth specimens was 11.8%.
We have investigated the effect of the notch inclination angle (φ), which is defined as the angle between the center line of a specimen and the center line of V-groove having notch opening angle of 90°, on the plane bending fatigue behavior. This study was conducted to reveal the influences of φ on the plane bending fatigue limit, the fatigue crack initiation and propagation and the behavior of non-propagating crack. The main results obtained are as follows; (1) The fatigue limit was increased and the fatigue life was prolonged with the increase of φ. Especially, the test specimen having the notch shape like a step (in this case, φ=45°) showed the tendency that the fatigue life is much longer than other cases (φ=0°, φ=15°, φ=30°). (2) In case of stress level having the inflexion point in the fatigue crack growth curves, the results obtained were independent of the influence of φ. The fatigue crack growth rate (dl/dN) decreased with an increase of crack length until it reached to the inflexion point of dl/dN in the fatigue crack growth curve. Then it increased or the fatigue crack became non-propagating, depended on the level of the applied stress. The similar tendency was found in the relation between dl/dN and the stress intensity factor range (ΔK), when ΔK was taken as a parameter instead of the crack length. In the increasing range of dl/dN, the following equation was experimentally obtained between dl/dN and ΔK. dl/dN=5.92×10-11×ΔK4.19 In the decreasing range of dl/dN, the relation between dl/dK and ΔK was able to be arranged using the stress concentration factor (α).
In order to clarify the cause for statistical scatter of fatigue lives, crack propagation behavior was investigated. Three kinds of metals, a low carbon steel (S15C) with four grain sizes, a pure copper and an austenitic stainless steel (SUS304), were tested under rotating bending to examine the effect of material structure. To evaluate the characteristics of scatter of crack propagation rate, the relationship between the crack propagation rate da/dN and the stress intensity factor range ΔK was analyzed for four ranges of crack length from crack initiation to failure. Statistical properties of a power m and a coefficient C in the Paris' law obtained by a regression analysis for each crack were examined in relation to material structure. Furthermore, the crack deflection ratio, Rd, obtained from the measurement of crack length along crack path, which increases with the effect of material structure, was statistically examined and discussed.
The continuous distribution of dislocations was used to model the slip deformation ahead of the tip of a small fatigue crack interacting with grain boundaries. The propagation rate of a fatigue crack was assumed to be controlled by the crack-tip opening displacement. A Monte Carlo simulation was conducted for fatigue crack growth by giving the grain size and the frictional stress of each grain as random variables following the two-parameter Weibull distribution. The distributions of the crack propagation rate and the crack propagation life were computed as functions of the crack length and the stress level as well as of the grain size. The results of simulation agreed at least qualitatively with the experimental results obtained for aluminum alloy 2024-T3.
Fatigue experiments were conducted to elucidate the influence of inclusions on fatigue crack growth behavior of high manganese steel at low stress amplitude near the threshold region and in the region II for the temperature range of 297K to 113K. Two series of steel with different inclusion contents were used. Compact tension specimens were prepared in SL or ST orientation with respect to the rolling direction. The results showed that the fatigue crack growth rate da/dN of SL specimen was higher than that of ST specimen for both series at low stress levels near the threshold region. On their fracture surfaces, slightly larger roughness was measured in the case of ST specimen compared with the SL specimen and the roughness increased slightly with an increase of the content of sulphide inclusion regardless of experimental temperature. Based on these results, it is considered that da/dN in this region is affected by the crack closure due to the surface roughness which depends on the inclusion content and on the orientation. In the region II, the influence of orientation on da/dN was more pronounced than that of content at both 297K and 113K. However, ST specimen with a high inclusion content showed the lower fatigue crack growth rate than the base material at 113K. In order to make clear this reason, fracture surface were observed by a SEM. It is pointed out that microcracks produced in elongated MnS inclusions increase da/dN for SL specimen but reduce da/dN for ST specimen.
The purpose of this study was to estimate the influence of amplitude fluctuation on fatigue damage. For this purpose, a new method to measure internal and effective stresses only from the stress-strain hysteresis data was developed. By using this method, the transient behaviours of total strain, plastic strain width and effective stress were observed during two-step stress cycling tests. Using a new cumulative damage criterion named ES-Miner's rule, fatigue damages under multiple repeated stresses in two levels were calculated and compared with those obtained by Miner's rule. The fatigue damage caused by stress amplitude fluctuations was evaluated on the viewpoint of internal stress and effective stress. The results obtained are as follows: 1. The influences of stress amplitude fluctuation on plastic strain width and effective stress leasted up to 1000 cycles. 2. The stress amplitude fluctuation, regardless of its direction, gave a transient damage greater than that under a constant stress amplitude.
The Correlation between behavior of small fatigue cracks and variation of cumulative fatigue damage under the multiple repeated two-step stress loading was investigated by using specimens with an initial small crack. Various combinations of the overstress σH and the understress σL cause various values of cumulative fatigue demage. The critical stress σwi for fatigue crack initiation from a small crack is the important measure to evaluate the contribution of σL on fatigue damage. The values of cumulative fatigue damage D based on the Modified Miner Rule became much smaller than 1.0 for a certain combination of σH and σL. The minimum value of D obtained in the present study was D≅0.18. Although a wide scatter of D has been reported by many studies, it is not caused by statistical scatters of a material but caused by various combinations and sequence of σH and σL in relation to crack size at every moment. The possible minimum value of D (Dmin) for a definite cycle ratio of σH and σL under various combinations of sequence of loadings can be predicted by the method proposed in the present study and Dmin will be useful for the design under variable amplitude loadings.
In order to make clear the effect of residual stress on fatigue crack growth behavior, fatigue crack growth tests under constant amplitude and repeated two-step loadings were carried out on SM50A steel in the near threshold region by using the specimens with or without longitudinal residual stress. The crack length and crack opening point were measured by the unloading elastic compliance method. The results obtained are as follows. (1) Under varying loadings, a fatigue crack was found to grow in the residual stress field even when the low level effective stress intensity range (ΔKeff)L was less than (ΔKeff)th. (2) However, the fatigue crack growth rate under low level load (da/dn)*L was influenced by the number of cycles of low level load NL. When NL was rather large, the crack growth increment ΔaL became saturated and the averaged growth rate decreased. (3) The modified Miner type linear accumulation law of crack growth in terms of ΔKeff was confirmed to give a conservative estimation of near threshold fatigue crack growth rate under varying loadings even in the residual stress field.
Fatigue crack growth tests were made in air and in 3.5% saltwater on 7075-T6 aluminum alloy specimens extructed in LT and TL directions. The testing in saltwater was made under the application of anodic and cathodic potentials. The test results showed that the crack growth rate in saltwater was higher than that in air under all the conditions tested, regardless of the applied potential and the specimen orientation. Although an appreciable effect of the applied potential was observed in the lower ΔKI range of the da/dN-ΔKI relation for LT specimen, almost no effect of potential was observed in both the da/dN-ΔKI relation for TL specimen and the da/dN-ΔKIeff relations for LT and TL specimens. Intergranular fracture and secondary cracking of intergranular type were observed quite often in the LT and TL specimens fatigued under free corrosion.
Fatigue properties of aluminum alloy reinforced by SiC particles were studied under rotating bending conditions. The tests on three kinds of SiC volume fraction (0%, 10%, 20%) were carried out, and both fatigue life and fatigue crack growth were observed. Furthermore, the causes of fracture and the fracture mechanism were investigated by an electron microscope. The results obtained are summarized as follows: (1) The fatigue limit of aluminum alloy with SiC particles became higher as the volume fraction increased. For example, the fatigue limit with 20% volume fraction rised up to 40% than that with 0%. (2) The influence of large particles of SiC, over 80-100μm, on the fatigue life was clarified. (3) The fatigue crack growth rate decreased with increasing SiC volume fraction, and the growth rate with 20% volume fraction was about a fifth of that with 0%.
In order to know the influences of load wave form and compressive loads on the crack propagation behavior of sintered Si3N4 under cyclic loads, the crack propagation rate was measured using compact tension specimens at room temperature. The main results obtained are as follows. (1) The influences of stress ratio and test frequency to the crack propagation behavior under sine wave form cyclic loads were equal to those under square wave form cyclic loads. (2) In the range of large KImax, the crack propagated nearly time dependently. In the range of small KImax, the crack propagated cycle dependently. (3) Crack closure phenomenon was observed using the unloading elastic compliance method. (4) The effective stress intensity factor range ΔKeff, together with KImax, can be a reasonable parameter to describe the crack propagation rate. (5) The compressive load crushed the debris, lowered the crack opening load and resulted in a large ΔKeff which promote the increase of the crack propagation rate.