The effect of stress triaxiality on ductile fracture was investigated on several structural steels. Round bar tensile specimens with circumferential notches of different acuity were used for the tests to develop the triaxial state of stress. The steels used were three mild steels, two medium strength steels and a 21/4Cr-1Mo low alloy steel. For all the steels tested, the size of dimples increased and the strain at the ductile crack initiation sharply decreased with an increase of stress triaxiality of the specimen. The degree of this stress triaxiality effect differed largely with materials. This material dependence of triaxiality effect may, mainly, be due to the difference in size distribution of inclusions. To investigate the effect of stress relief annealing on void formation and growth, the tests, in which annealing after 20% (or 40%) straining was repeated until fracture, were also carried out. The test results showed that, in the notched specimens, the voids grew to huge dimples by the repetition of annealing after 20% straining.
An instrumented impact charpy test was carried out to evaluate the effects of the notch shape, hardness, and carburized depth on the impact bending strength of notched high strength steels (SCM 440 and SCM 415 steels). The principal results were as follows: (1) The impact bending strength of the notched high strength steel was in near proportion to the tensile strength of the material, having no direct corresponding relationship with the absorbed impact energy. (2) However, the impact bending strength showed an ultimate value which is determined by the notch shape and the tensile strength of the material, and a proportional relationship existed between the impact bending strength and tensile strength below this ultimate value. (3) The impact bending strength of the carburized SCM 415 steel showed a value of about 10% larger than the static bending strength. On the other hand, the impact bending strength of the SCM 440 steel showed a value of about 10% smaller than the static bending strength, and this tendency was contrary to the general trend. (4) The result of fractographic observation by SEM showed that the width of the DIZ (ductile initiation zone between notch root and unstable fracture area) depended largely upon the notch shape and that only the data at such high strength levels as HRC=53 and 55 were in proportion to the impact bending strength.
Impact tensile tests of tufftrided low carbon steel were carried out at various temperatures by using a drop-weight type impact testing machine designed in our laboratory. The initiation of crack during the fracture process was observed by using a stroboscope. The fractography and the observation of microstructure near the crack tip were performed by optical and electron microscopes. The transition temperature rose by tufftriding, and this rise became more remarkable with increasing treating time. This can be explained by the increase of yield stress and the decrease of brittle fracture stress. At the high temperature side of transition region in tufftrided steel, brittle cracks initiated near the surface of hardened layer and propagated to the core. The fracture mode changed from brittle fracture (cleavage) to ductile fracture (dimple) at the boundry between the case layer and the core. By considering these characteristics, a fracture model of tufftrided steel was proposed and the fracture mechanism was discussed.
Fracture behavior of the weld zone of 80kg/mm2 class HSLA steel has been investigated by means of tensile test, Charpy impact test and three point bending test. The results obtained gave convincible evidences that showed the following important role played by the martensite-austenite (M-A) constituent on the initiation and propagation processes of cracks. For ductile fracture, cracks were initiated by cracking flaky M-A constituent or debonding blocky M-A constituent from ferrite matrix. Then cracks expanded to voids and developed to deep holes. The main crack was formed by coalescence of deep holes by internal necking and led to fracture of specimen with a ductile deep dimple fracture surface. On the other hand, for brittle fracture, cracks nucleated at the interface of blocky M-A constituent and ferrite matrix. Then the crack propagated into the specimen, resulting in catastrophic fracture with a brittle river pattern surface.
Three-point bending fracture tests were carried out on three kinds of high alumina ceramics. The specimens were partially cracked by thermal shock before the fracture tests. The test results could be better evaluated in terms of nominal fracture stress, σc, using the effective transversal area, rather than the critical stress intensity factor, KIc. The results evaluated through KIc depended on the pre-crack length. However, σc's were almost independent of the crack length. Little topographic difference between the thermally-shocked and mechanically fractured surfaces was observed. For pure alumina specimens consisting of coarse grains, transgranular cracks were mainly observed on the thermally pre-cracked surface, while intergranular cracks were major on the mechanically fractured surface. Acoustic emissions were also monitored, and those signals were recorded using a transient memory recorder. The magnitude of the signals was large for porous mullite ceramics and small for fine-grain alumina ceramics. For mullite ceramics, the fracture stress obtained from cyclic loading was about 30% higher than that from monotonic loading. Cyclic step patterns with about 10 to 20μm spacings were observed on the cyclic fracture surface.
Computer image processing technology has been applied to fracture surface analysis, and the generalized software for 3-dimensional fractography image analysis has been developed. By this software, it became possible to obtain 3-dimensional fracture surface topography from a stereo-pair of SEM fractograph with a sufficient precision. The fatigue fracture surface roughness of a hightensile strenght steel HT 55 in air and in vacuum was calculated, and the root mean square roughness of fracture surface in vacuum was made clear to be greater than that in air at R=-1, compared with those at R=0.1 and 0.5. The wedge effect induced by corrosion products was also clearly demonstrated on a matching-pair of corrosion-fatigue fracture surface topography of HT 55.
Thin plate specimens with 1mm thickness were cut out from the different micro structures of a large casting and subjected to fatigue tests under a condition so as to minimize the effects of defects such as porosity, shrinkage and non-metallic inclusions. The influence of casting structure on fatigue behavior was experimentally examined through these fatigue tests and fractographical investigation on the crack initiation and propagation behavior of steel castings. The results of fatigue life tests on the unnotched and V notched specimens showed that the fine grain material had a higher fatigue limit as compared with the coarse grain material and the influence of cast structure on crack initiation was observed. On the other hand, the influence of micro structure on the crack propagation rate was not observed above the crack propagation rate of 10-6mm/cycle. In the crack initiation area of the fine and coarse grain materials, cast structure sensitive fracture surface was predominantly observed. Striation appeared clearly at K>12.4MN·m-3/2. The crack propagated preferentially in the ferrite phase and slowed down when the crack tip met with the pearlite phase.
Pulsating tensile fatigue tests were carried out, using the CT specimens cut out from a rolled steel having remarkable laminated structures, in three different directions. The crack growth rate in each directional specimen was measured and the fracture surface was observed by a scanning electron microscope. The main results obtained are as follows: (1) An anisotropy of crack propagation was observed in a severely laminated structure, even in the case that inclusions hardly contributed to crack propagation. (2) The resistance for fatigue crack propagation was smallest in the specimen (SL specimen) whose cracks propagated along the laminated structure, and it was largest in the specimen (LS specimen) whose cracks propagated by cutting the laminated structure alternately. (3) The existence of the anisotropy of crack propagation was confirmed also by the observation of fracture surface. That is, in the specimen (LS specimen) whose resistance for fatigue crack propagation was largest, the secondary cracks originating from the laminated structure were observed frequently at right angle to the direction of crack propagation.
The mechanism of the initiation of fish eye failure was examined by stereography of the fracture surface of SCr 420 carburized steel tested under rotating bending. Three modes of fish eye nucleation were observed: mode I crack initiation from nonmetallic inclusion (ni), mode II crack initiation from ni followed by the transition to mode I crack, and mode II crack initiation at the martensitic lath interface. The mode II crack plane was close to or less inclined than the plane of maximum shear stress. The maximum length of mode II crack was longer when the local stress at the origin was lower. The transition from mode II to mode I occurred at the threshold mode I stress intensity factor, KIth, ≅3MPam1/2. The threshold mode II stress intensity factor for the nucleation was estimated as KIIth≅1.5MPam1/2.
Fatigue crack growth behavior and fracture toughness of a structual steel, JIS SM50A, have been investigated at room temperaure and 123K, and the influences of load ratio, R, and frequency on the fatigue crack growth rate were examined. Low temperature was achieved by changing the flow quantity of liquid nitrogen. An electron fractographic analysis was employed to determine the mechanism of fatigue crack growth. At 123K, the fatigue crack growth rate was considerably affected by the load ratio, and crack growth curves indicated a marked acceleration to high ΔK regions near the final failure, because of cleavage during the fatigue crack growth. On the other hand, the crack growth rate at room temperature was found to be insensitive to the load ratio where the mechanism of growth does not involve a static fracture mode. At 123K, the beltlike traces of cleavage appeared on the fatigue fracture surface as the evidence of cyclic cleavage when the fatigue crack exceeded the value of Kmax about 25MPa√m. The value of Kci at which the onset of cyclic cleavage occurs was not influenced by the load ratio, frequency or loading condition, and the Kci value was equivalent to about 70% of the value of fatigue fracture toughness, Kfc. The cleavage width on the fracture surface tended to increase with increasing Kmax. Formation of cyclic cleavage may be related to the local embrittlement caused by work-hardening in the plastic zone ahead of the fatigue crack. Kfc was slightly decreased as compared with the value of Kc obtained by monotonic tensile loading. Striation spacings were found to be reduced when temperature was decreased. However the striation spacings were closely correlated with the effective stress intensity factor range, ΔKeff, regardless of temperature.
The phenomenon of fatigue crack growth acceleration due to variable amplitude overloading was studied in a 2017-T3 aluminum alloy. Compact specimens were subjected to single and multiple peak applications of overloads. Striation spacings during and just after the overloads were measured. The crack tip blunting and resharpening process for each loading cycle was examined. The fully blunted crack tip was followed by fatigue crack growth acceleration microscopically. On the other hand, no fatigue crack growth acceleration occurred if the previous crack tip was fully resharpened. These results were able to be interpreted from the view point of the transition from the fatigue crack to the ideal crack. It should be noted, however, that the shape of the crack tip during overloading had no marked influence on the subsequent fatigue crack growth retardation macroscopically.
For prevention of fatigue fracture troubles in structural parts, the establishment of a procedure to evaluate service load by micro fractography is desirable. Fatigue tests under block loading with four or five load steps and random loading with a constant minimum load were carried out using CT test specimens of stainless steel (SUS 304). Micro-fractographic analysis was performed on the test specimens and the relationship between the applied stress and the striation spacing observed was investigated. The results obtained are summarized as follows: (1) The striation spacing corresponding to the maximum stress amplitude in the block loading tests was equal to that in the constant load test. (2) Retardation of fatigue crack propagation was recognized in the striation spacing corresponding to the stress amplitude lower than the maximum amplitude in the block loading test. (3) The data of fatigue crack propagation rate obtained in the block and random loading tests fitted well to a straight line when plotted against m factorial mean ΔK value. (5) The random test loads could be evaluated almost correctly from the striation spacings measured, except for small load spectra which gave unresolvably small striation spacings.
Propagation of microscopic fatigue cracks by periodic overstressing was studied under different overstress and understress conditions using a low carbon steel. When the overstress value was large, a significant acceleration of crack propagation (more than one hundred times) occurred even if the understress value was lower than the threshold stress, similarly as in the case of macroscopic cracks. The lower limit of the understress that caused such a significant acceleration was reduced as the overstress was increased, but it was independent of the crack length as far as the crack was microscopic (less than 50μm). For longer cracks, the lower limit was considerably reduced by increasing the crack length. When the overstress value was low, a significant acceleration in short cracks occurred only when the understress value was nearly equal to the threshold stress, but in longer cracks it took place even the understress value was lower than the threshold stress. The microscopic fracture surface consisted of small facets, which was supposed to be related to crystal structures. Observation of the specimen surface showed that cracks propagated intermittently, hesitating at the boundary of microstructure. These observations indicate that microstructure has a significant effect upon microscopic crack propagation under intermittent overstressing.
Fatigue crack growth in SUS 304 stainless steel was investigated at 650°C under trapezoidal and sinusoidal stress waves. Special attention was placed upon the interrelation between the high-temperature fatigue crack growth behavior and fractographs. At high and intermediate crack growth rates, ΔJf (J integral range)-controlled cycle-dependent crack growth mode and J' (modified J integral)-controlled time-dependent crack growth mode were observed, which yielded transgranular and intergranular cracking, respectively. At low crack growth rates, a threshold of ΔJf was found to exist, below which the transgranular cycle-dependent crack growth mode vanished and the intergranular time-dependent crack growth mode emerged. This generated false fatigue thresholds, the levels of which were lower than the true one and arbitrary depending on the condition of stress cycling. In ΔJf (or stress intensity range ΔK) decreasing tests, fracture mode changed from transgranular to intergranular, in accordance with the transition from the cycle-dependent crack growth to the time-dependent one, which occurred when ΔJf crossed the true threshold.
SCC susceptibility of sensitized SUS 304 stainless steel in high temperature water was studied. The results obtained are as follows. SCC susceptibility was increased by adding crevices to the tensile specimen surface, for the corrodent became acidified by hydrolysis in crevices. SCC susceptibility was best fit to TTS curve obtained by EPR test, not by other corrosion tests such as Strauss test or the grain boundary corrosion test in high temperature water. In addition, by giving a simulated weld thermal cycle before the sensitizing heat treatment, the sensitization was clearly promoted. This seemed to be caused by the reason that nucleation of carbide occured in the simulated weld thermal cycle process and it promoted the carbide growth and the formation of Cr poor layer around carbide in the subsequent sensitization process.
The relation between hydride precipitation behavior and fracture morphology was investigated on Ti-6Al-4V alloy by means of fractography and internal friction measurement. When the specimen hydrogenated thermally at 1273K for 7.2ks was furnace-cooled (β annealed specimen), fine particles of hydride precipitated from the β phase during cooling. The fracture morphology showed the occurance of fracture at the interface between the β phase and the hydride or inside the hydrides. The strength of the hydride was 0.5GPa and smaller than that of the matrix, and the elongation was hardly present. The hydride acted as the fracture initiation site. When the same hydrogenated specimen was water-quenched (β quenched specimen), the hydride hardly precipitated and hydrogen was solutionized compulsorily in the β phase. When the β quenched specimen was plastic-deformed, the hydride precipitated from the β phase. At a high tensile strain rate, the fracture morphology was similar to that of the β annealed specimen. This means that the hydride precipitated finely from the β phase by plastic deformation and the fracture occurred at the interface between the hydride and the β phase or inside the β phase. At a low tensile strain rate, the fracture morphology showed the terrace fracture inside the hydride. This means that the hydride precipitated from the β phase, grew to a large block independent of microstructure, and then fracture occurred inside the hydride.
The fracture toughness and the fatigue crack growth properties of heat treated SNCM439 steel were investigated under high pressure of hydrogen up to 9.9MPa at room temperature. The results obtained are as follows; (1) The fracture toughness, KQ value, of the steel in hydrogen was smaller than that in argon independent of heat treatments. KQ values of the quenched and the tempered steel were smaller than that of the annealed steel in hydrogen. On the fracture surface of these steels, the streched zone was not observed in hydrogen but in argon. (2) The fatigue crack growth rate of the steel in hydrogen was larger than that in argon, and KFc of the steel in hydrogen was smaller than that in argon independent of heat treatments. The order to hydrogen susceptibility of fatigue crack growth properties increased from the annealed to the quenched and then to the tempered steel. (3) The quasi-cleavage fracture did not occur in argon but in hydrogen on the fatigue fracture surface of the annealed steel. Both the amount of the intergranular fracture and the surface roughness of the transgranular fracture in hydrogen were larger than those in argon on the fatigue fracture surface of the quenched and the tempered steels.
The low-cycle fatigue tests at various strain rates and strain wave forms were conducted on sensitized and solution treated 304 S.S. in 573K water with 8ppm dissolved oxygen. The results based on the fractographical study are as follows: (1) The low cycle fatigue life of 304 S.S. in high temperature water was shorter than that in air and affected by sensitization, strain rate and strain wafe form. (2) In the sensitized 304 S.S., cracks initiated in intergranular mode, while in the solution treated one they initiated in transgranular mode. (3) In the sensitized 304 S.S., the fracture mode changed from the intergranular type to the transgranular type when the crack propagation rate (da/dt)T reached to about 1×10-8m/sec. (4) When cracks propagated in transgranular mode, the crack propagation rate was affected by the tension going frequency, not by sensitization.