PSI - Issue 23

M.A. Artamonov et al. / Procedia Structural Integrity 23 (2019) 257–262 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

261

5

Fig. 3. Dependence of (a) total lifetime of specimen, (b) m coefficient from Paris equation, (c) number of crack growth cycles and (d) duration of the fatigue crack initiation on temperature.

Discussion

The kinetics of temperature and cycle asymmetry influence on the period of crack initiation demonstrated that the great effect on the crack initiation is exerted by the maximum level of strain in the cycle but not by the magnitude of strain. Therefore, when the cycle asymmetry is equal to R = 0.5, the crack initiates much faster than at R = 0. The increase in temperature brings about the decrease in the period of crack initiation, at that the dependence of N o on T is well approximated by the linear dependence (the correlation coefficient is r ≈ 0.74 - 0.84). The dependence of durability on temperature is more complex. For the asymmetry ratio of R = 0 one can observe a characteristic feature: when th e temperature changes from 20 °C to 550 °C, the durability of specimens does not decrease, in fact there is even a slight increase. With the further increase in temperature, the durability of specimens sharply decreases (Fig. 3a). In order to explain this phenomenon, it can be noted that the durability of specimen depends not only on the FCG rate, determined by the Paris parameters (m and C), but also on the time of the crack growth stability. The increase in temperature increases the plasticity of material and delays the transition from the stable FCG to the accelerated growth. For the asymmetry ratio of R = 0.5 the durability demonstrates better results in comparison with the asymmetry ration of R = 0. This is especially appears at low temperatures. Apparently, the higher plasticity, which is attended at these test conditions, leads to an increase in the duration of FCG. The results of the Paris m coefficient calculation, obtained at the asymmetry ratio of R = 0, demonstrates the same nature of dependence on temperature as for the m values obtained by the conventional method on the CT specimens (Fig. 3b) [12]. At high temperatures one can observe the deviation in m factor between the cylindrical specimens and CT specimens. Perhaps, this difference in the nature of the crack growth dynamics is associated with the influence of two factors. For the cylindrical specimens the shape of fatigue cracks is close to the elliptical form, while for the CT specimens the front of fatigue crack is linear [13]. The second factor is the different type of mechanism of the fatigue failure during the tests: for the cylindrical specimens - testing according to the LCF, for the plain CT specimens testing is generally carried out according to the mechanism of high-cycle fatigue. The SIF magnitude during the crack growth may be similar, but the cylindrical specimen begins to change already in the first loading cycle, and as the fatigue crack propagates in the material its properties may differ from the initial ones.