13th International Conference on Fracture June 16–21, 2013, Beijing, China -5mechanism in both specimens. This trend can be seen at 673 K. However, since the value of n indicates the morphology of precipitates or the precipitation process [11], the smaller values of n at 673 K suggest that the nature of precipitates at this temperature may differ from that of 473 K. It should be also noted that the evaluated incubation time is very small except SA-P specimen aged at 673 K, suggesting that the precipitation starts very quickly. The assumed activation energy of pipe diffusion, Qd, can be evaluated from the values of β10T. Fig. 4 shows the Arrhenius plots of β10T evaluated for SA-P and SA-U0 specimens. The slopes of straight lines in this figure yield 82.9 kJ/mol and 113 kJ/mol for the value of Qd in SA-P and SP-U specimens, respectively (Table 3). These values of Qd are nearly one third or half of the activation energy of lattice diffusion, Ql (for instance, Ql = 244 kJ/mol for Ni in pure Fe), and close to the activation energy of dynamic strain aging in Fe-Ni alloys (~110 kJ/mol) and Fe-Si alloy (125 kJ/mol) measured by Cuddy and Leslie [14]. Cuddy and Leslie suggest that the pipe diffusion of substitutional solute atoms which have large solution-hardening effects contributes to the occurrence of dynamic strain aging, since C atoms in the alloys they used are believed to be removed by the addition of Ti. The contribution of C atoms is also believed to be negligibly small in the maraging steels investigated here, because the content of C atoms is too small to yield the observed increase of hardness and the diffusion coefficients of C atoms at 473 K and 673 K are too large to explain the very slow increase of hardness at these temperatures. 3.3 Dynamic aging during fatigue tests Fig. 5 shows the SNcurves of SA-U0 specimen tested at room temperature, 473 K and 673 K. The increase in test temperature markedly increases the fatigue strength in high cycle fatigue region, and the increase of fatigue strength at the number of cycles larger than about 5 x 105 seems to be larger at 673 K than at 473 K. These results are in contrary with the fact that the yield strength decreases as the test temperature is increased [5]. In order to reassess the previous data, we use the following formula for the relation of stress amplitude ( σa) with the number of cycles (Nf). σa = σao +k{(Nf/Nfo) −a −1}b. (4) Here Nfo is a reference number of cycles to failure, σa is the fatigue strength at the reference number, k is the strength factor, and a and b are numerical constants. Since the fatigue fracture does not take place at cycles larger than 107, as is shown by arrows in Fig. 5, the value of Nfo is set to be 107. The fatigue limit defined by σao is evaluated to be 597 MPa, 745 MPa and 796 MPa at roomtemperature, 473 K and 673 K, respectively. In Fig. 6, the abscissa of Fig. 5 is converted into the time to failure tf and the extrapolation to low cycle fatigue region is conducted by using Eq. (4). The σa-tf curves of Fig. 5. SN curves of 300G steel tested at Fig. 6. Relation of stress amplitude with time room temperature, 473 K and 673 K. to failure in 300G steel. 400 600 800 1000 1200 104 105 106 107 RT 473 K 673 K Stress amplitude σ a (MPa) Number of cycles to failure Nf 300G (SA-U0) 400 600 800 1000 1200 1 10 100 RT 473 K 673 K Stress amplitude σ a (MPa) Time to failure tf 300G (SA-U0) (ks)
RkJQdWJsaXNoZXIy MjM0NDE=