Fracture toughness tests and unloading tests of C-1/2Mo steel and 21/4Cr-1Mo steel were conducted on CT specimens with different thickness at the ductile-brittle transition temperatures. According to fractographic examination of microscopic ductile tearing at the crack tip on the fracture surface of the specimens by a scanning electron microscope, the following two test procedures were proposed to estimate the plane strain fracture toughness KIc of small size specimens at the transition temperature; (1) The modified JIc test procedure to determine the value of J or KI at the initiation of microscopic ductile tearing at the midsection of the specimen which has critical volume as a trigger of cleavage fracture, and (2) the interpolation technique using the values of KIc which are determined by JIc test at the upper shelf temperature and KIc test at the lower shelf temperature.
Since specimens taken from weldments or thermally simulated samples have a size limitation, a need has arisen to measure fracture toughness by using Charpy-size small specimens. In this study, JIc fracture toughness tests using two types of specimens (compact and three-point-bend specimens) with various dimensions were conducted for medium strength steels by the unloading compliance method in order to determine the applicability of Charpy-size specimens. Also, the metallurgical effect on fracture toughness of the heat-affected zone was discussed by the JIc test results of Charpy-size thermally simulated specimens. The following conclusions were obtained: (1) By using a side-groove to 20% of the original thickness of the specimen, J-resistance curves independent of specimen configuration were obtained. (2) Although Charpy-size specimens with side-groove did not satisfy the size requirement of JIc validity (B, W-a≥25JIc/σf), they had the same critical J value at the crack initiation as that of the larger-size specimens. (3) JIc fracture toughness of the heat affected zone can be expressed very well by the experimental equation JIc=(-AσB+B)/S where A=0.152, B=11.7, σB: tensile strength in kg/mm2 and S: area fraction of nonmetallic inclusions in percent.
The forging process of large products sometimes requires more than a month for repeated heating to high temperatures. The heating of steels up to an extremely high temperature during forging causes precipitation of sulfide inclusions at the grain boundaries which makes the grain boundaries weaken. The phenomenon is called overheating which sometimes reveals a loss of ductility and toughness in tensile and impact tests. In this study, the effect of overheating on the fracture toughness was investigated using Ni-Cr-Mo-V steels with various sulfur contents. It was revealed that the overheating led to a decrease in fracture toughness JIc and tearing modulus TJ of the materials, depending on sulfur content, heating temperature and cooling rate. The degradation was caused by the crack initiation and propagation through the grain boundaries weakened due to the manganese sulfide precipitates. Quenching and tempering heat treatments improved the toughness properties of the overheated steels.
In large forgings such as turbine rotors, it is often recognized that small clustered flaws or heavy sulfide inclusions are scattered in the center part. These flaws or inclusions are too small to be evaluated individually by the linear elastic fracture mechanics (LEFM) and their scattered area is too large to be evaluated for understanding unstable fracture of the forgings. The present study was made to find out a simple way to evaluate these flaws by using the fracture toughness of matrix (KIc, JIc) and simple bending fracture tests of smooth specimens containing flaws. An equivalent crack (Aeq) was calculated from these tests by different methods. The equivalent single crack calculated by using J-integral was the same order as individual small flaws, while that calculated by LEFM gave the same order as the sulfur segregation band width. JIc seemed to express the initiation of unstable fracture from individual small flaws. A simpler value “f”(=√mAeq) instead of the equivalent crack (Aeq) was proposed for the evaluation of these clustered flaws.
The standard of Japan Welding Engineering Society WES 2805 gives the procedure for safety evaluation of defects in welded structures on the basis of the COD concept. The establishment of an appropriate relationship among applied load, defect size and COD value, so-called.“COD design curve”, is essential in the application of the COD concept to the safety evaluation of structures. The COD design curve adopted in WES 2805 takes the form δ=3.5ea Where e: applied strain a: equivalent through-thickness flaw size parameter In practice, attention must be paid particularly to brittle fracture initiation from a part-through crack in a highly strain concentrated region. For practical use of the COD design curve, experimental verification must be made about the relevancy of the LEFM-based equivalency of flaw size parameter a for a general yielding condition, and also about the evaluation method of applied strain e for a localized highly strained region. In this study, bending tests using surface-cracked specimens were carried out, and it was shown that the determination procedure of parameter a currently used in WES 2805 is reasonable even in the elastic-plastic regime. Furthermore, from the experiments using tensile specimens, in each of which two corner cracks existed in stress concentrated region, a way of defining the applied strain e in the highly strained region with steep strain gradient in it and a new COD design curve were proposed.
The effects of stress triaxiality and temperature on the ductile fracture process have been investigated with particular reference to the micro-void formation and growth. The steels used were SM50A, SM50B and 21/4Cr-1Mo-steel. To develop triaxial stress of various acuity, round bar tensile specimens with circumferential notch of different acuity were chosen for the tests. From the metallographic observation on the sections of the specimens off-loaded on the way of loading, it was found that both the formation and the growth of voids were significantly influenced by stress triaxiality. Temeprature mainly affects the void formation process rather than its growth. The effect of temperature may be assumed to be caused by its effect on the flow stress of the material. Fracture toughness tests on CT specimens have also been made in order to obtain the relation between the fracture toughness and the ductility under the axisymmetric condition with the high triaxiality.
In nuclear power plants, it is an important problem to estimate the structural safety margin of a reactor vessel under a pressurized thermal shock, which may occur in an accident such as small break loss of coolant. Therefore, from the view point of fracture mechanics, this paper deals both experimentally and numerically with the behavior of cracks existing in the A508 class 3 pressure vessel steel plate under the combined load of thermal shock and tension, which simulates the above mentioned phenomenon in an actual vessel. In order to estimate simply the structural integrity of the nozzle corner as a special case, the initial cracks located at the edge of the hole in the specimen were considered. To take into account the influence of the compliance of the pressure vessel itself and pressurized water, the pre-load was mechanically applied with a spring which was set in series with the specimen. By holding the pre-load at the initial elevated temperature, the specimen was quenched with cold water rapidly. The experimental results showed that the initial cracks extended unstably due to the thermal shock loading after a stable extension of a few milimeters.
With an increasing use of polymeric materials in structural components, fracture prediction of these materials under various loading conditions is becoming of great importance. The authors reported previously that the crack-tip singular stress field of a viscoelastic body was characterized by three parameters: the stress intensity factor KI(t), the strain intensity factor TI(t), and the energy release rate J. In this paper, the results of impact fracture tests on PMMA specimens are presented to examine the applicability of these parameters. The three-point bend specimens were fractured at 20°C by a falling steel sphere or cylinder over the range of loading rates 103MN/m3/2s≤KI(t)≤104MN/m3/2s. The fracture initiation time was measured by a strain gauge mounted near the crack-tip or a painted electrical resistance circuit in the path of the crack. The contact force between the specimen and impactor was analyzed by applying Hertz's theory to the local deformation near the contact point, and the fracture mechanics parameters were determined by using the simple formulas which had been derived by the authors using Timoshenko's beam theory. The critical values of the fracture mechanics parameters at crack initiation were compared with the quasi-static results reported in the previous paper. It was found that the critical values of KI(t) and J increased with loading rate, while the critical value of TI(t) was constant over the range of loading rates 10-3MN/m3/2s≤KI(t)≤104MN/m3/2s. An examination was also made on the fractography of the specimens, and the critical value of the crack opening displacement was found to be constant. This fact is consistent with the above experimental results, since the strain intensity factor criterion is equivalent to the crack-tip opening displacement criterion.
Closure behavior and growth characteristics of a wide range of fatigue cracks including incipient short cracks and elastic-plastic long ones extending beyond general yielding were investigated by using the unloading elastic compliance method, and in situ direct observations of propagating fatigue cracks were made with a scanning electron microscope to study their microscopic crack growth modes. The effective stress intensity range ΔKeff was found the most governing parameter of growth rate for all kinds of fatigue cracks, except when general yielding took place and monotonic deformation became significant. It was found that the microscopical crack growth mode varied depending on growth rate and the crack opening ratio U was related to growth mode.
Environmental fatigue crack growth tests of a high-tensile strength steel HT 55 have been conducted in dry air and vacuum, and crack closure behavior as well as fracture surface morphologies were investigated. The crack growth rate da/dN was significantly reduced in vacuum due to fracture surface rewelding. da/dN in vacuum at R=0.1 was equivalent to the one at R=0.5 in case of ΔKeff greater than 15MPa·m1/2, whilst the former became higher than the latter at ΔKeff less than 15MPa·m1/2. Roughness-induced unconformity of fracture surfaces is considered to prevent the rewelding of newly exposed surface at low R. At high R, on the other hand, no crack closure and no prevention of the rewelding were observed, thereby reducing the growth rate particularly near the threshold region. The crack opening value Kop rised immediately and the growth rate decreased rapidly with the environmental change from dry air to vacuum. On the contrary, a rapid increase of da/dN as well as a rapid decrease of Kop was observed with the environmental change from vacuum to dry air. Fracture surface morphology also varied abruptly with the environmental change. It is concluded that an absorption of gas molecules to newly created surface affects directly the crack closure behavior. Clay-like products induced by a remarkable plastic flow and rewelding were observed on the fracture surface in vaccum.
Fatigue crack propagation behaviour under the condition of mixed modes I and II applying was experimentally studied by applying tensile and torsional fatigue loadings of the same phase simultaneously on thin-walled hollow cylindrical specimens. As the initial component of ΔKII increased, the fatigue crack growth rate accelerated in the low crack growth rate region. In the rigion, where da/dN=A(ΔK)m holds, that is, in the second region of da/dN versus ΔK curve, the ratio of the initial ΔKII to the initial ΔKI, ΔKII/ΔKI, increased and reached near the value of 0.58, the fatigue crack growth rate became much lower than that at any other condition. When the value of initial ratio ΔKII/ΔKI increased further more, for example, ΔKII/ΔKI reached 2.16, the fatigue crack growth rate accelerated again. Under shear fatigue loading, the crack growth rate was higher than that of tensile fatigue loading.
The fatigue life prediction of butt welded joints was investigated by paying attention to fatigue cracks originating at the weld toes. The experimental results indicated that the fatigue crack initiation life depended on the weld toe configuration. The effect of stress concentration at the weld toes was estimated by a local strain approach with due consideration for the scatter of the stress concentration factor. The shallow fatigue cracks, 0.5-1.0mm, initially grew fast under the influence of stress concentration at the weld toes. The propagation of cracks more than 0.5-1.0mm was not affected by the weld toes. Since in this experiment, the fatigue crack initiation was defined as the state in which the crack depth became 0.5mm, this experimental result showed that the fatigue crack propagation life hardly depended on the weld toe configuration. From an engineering point of view, the fatigue crack propagation life can be estimated conveniently by the linear fracture mechanics analysis based on the assumption of the initial crack depth of 0.5-1.0mm at the weld toes, as described above.
In order to investigate the effect of the transition from small scale creep (SCC) to large scale creep (LSC) on the behavior of time-dependent fatigue crack propagation, stress-controlled fatigue crack propagation tests were carried out at 923K under trapezoidal stress waveforms on Inconel 718 in which such a transition was expected to occur. The results obtained were summarized as follows: (1) Grain-boundary facets were observed on the fracture surface in all cases. This indicates that the crack propagation proceeds by creep fracture. (2) The crack propagation rate, dl/dt, could not be correlated with the stress intensity factor. Also, the steady state creep J integral, Js*, was not a good parameter to represent dl/dt. These findings were associated with the effect of the transition from SCC to LSC on the crack propagation behavior. (3) As a new parameter giving a good correlation with dl/dt, the time-averaged creep J integral, Jt*, was proposed. It was defined as follows: Jt*=(∫tH0J*dt)/tH where tH: tensile stress hold-time, and J*: creep J integral under the transition from SCC to LSC. (4) The dl/dt vs. Jt* relation obtained in the time-dependent fatigue coincided with the dl/dt vs. Js* relation in static creep under the LSC condition.
Macroscopic through-crack propagation in ductile materials under creep conditions has been described successfully by the creep J-integral, J. This paper extended the creep J-integral concept based on the power law creep to the case of shorter cracks, and limitation of the use of J was examined. The propagation of surface cracks from a small semi-circular notch of about 0.6mm diameter and that from the surface grain boundaries of the smooth specimen with 40μm grain diameter of 304 stainless steel were observed during the air and the vacuum tests. When the crack grew to the length (depth) of 0.1mm or longer, the J approach became applicable, resulting in the same correlation between dc/dt and J as that for the macro-through-crack. Cracks less than 0.1mm, or two or three grain boundary length, on the other hand, showed higher propagation rates than that estimated by the straight line extrapolation on log-log plots of dc/dt-J. The growth of creep cavities by grain boundary diffusion was considered to be a dominant mechanism for the microscopic crack propagation. The time needed to make one grain facet fracture (crack initiation) at several surface grain boundaries can be well predicted on the basis of the model of cavity growth and coalescence proposed by Chen and Argon.
For the safety design of the pressure vessel of a reactor, the three dimensional J integral was evaluated by assuming a crack under thermal transient loadings and by using the finite element method. First, a semi-circular surface crack on an inner surface of the pressure vessel was analyzed. The thermal loading was assumed to occur by the injection of coolant. It was shown that the J integral increased with the time lapse gradually, but after some time they began to decrease. This phenomenon was due to the contraction of the vessel wall caused by a decrease in temperature. The effect of the mesh division on the path independency of the J integral was also studied. Secondly, a nozzle corner crack was analyzed using several crack shapes. In every case, it was shown that the J value became maximum near the nozzle surface. The superimposing of the J value by the inner pressure on those of thermal loadings was carried out and discussed.
Feasibility of DC potential technique for monitoring surface cracks was discussed by applying the reversing DC potential method to the plates with a simulated surface crack. Disturbance of electrical field occurred near the crack. Then the sensitivity of the crack measurement rapidly decreased with the distance of measuring position from the crack. Less than a half of the plate thickness was desirable as the measuring distance. Although the potential difference depended on the crack aspect ratio a/c, the crack area could be estimated from the potential ratio at the crack center. The potential distribution along the crack almost corresponded to the crack shape. A lack of singularity at the crack tip on the surface, however, made it difficult to determine the crack tip position. By using a proportional relation between the potential ratio V/V0 and the crack depth a/t, and by assuming that a surface crack takes a semi-circular configuration, the crack length c on the surface and the crack depth can be estimated. The optimum method to determine the crack shape is to repeat the finite element analysis until the potential obtained analytically coincides with that obtained experimentally.