PSI - Issue 5

A. Eremin et al. / Procedia Structural Integrity 5 (2017) 889–895

892

A. Eremin et al/ Structural Integrity Procedia 00 (2017) 000 – 000

4

2.2. Single Overload: measurement and loading procedure

The overload parameters. Basic cycles: asymmetry R =0.1, P max =2.8 kN, crack length a ≈ 14 mm, K max =13.91 MPa √m . The highest load at the overload cycle P OvL =4.2 kN. The overload plastic zone size for plane-stress conditions R y ≈ 680 µm. The loading sequence for the test 2 with the single overload and measurement cycles are schematically presented in fig. 2b. Actual steps are: 1. Precrack growth (at the basic load) was carried out before starting the measurement cycles till a crack length increment reaches 0.8 mm; 2. Application of 10 000 basic cycles to stabilize the crack growth rate and ensure linear direction of its propagation; 3. Application of two “slow cycles” at 0.03 Hz ( P =0.28-2.8-0.28 kN) and step-by-step image capturing at 0.3 fps. These reference measurements denote initial stage of the crack growth before running the underload cycle; 4. Application of the overload cycle with the gradual load increase from 1.82 kN up to 4.2 kN; 5. After passing the overload cycle image capturing is carried out at various time steps: immediately after the underload cycle, and then after 0, 100, 200, 500, 1000, 1500 and 2000 basic cycles. These are so-called “ measurement cycles ” . Measurements of the COD were performed by software extensometer placed in two locations – i) at the crack tip and ii) at the distance of 500 µm behind the crack tip (i.e. in the region where the crack wake has been already formed). 3.1. Single Underload Fig. 3 demonstrates hysteresis loops measured at the crack tip at different time steps – (a) for the first and the second reference cycles as well as at the cycle coming immediately after the underload; (b) after 0, 5 th , 10 th , 30 th , 50 th and 100 th basic cycles since the underload was applied. It is evident that the loop measured at the time step of “ 0 cycle ” substantially differs from the others. This is manifested in the higher value of ε and the loop area. Also, the loop is not closed. The observed gap between the starting and ending points of the loop (ε start =6.98·10 -6 and ε end =5.6·10 -4 values) is suggested to correspond to residual tensile strains at the crack tip which keep the crack opened and therefore drastically reduce K op . The loops measured at the following time steps are shown in Fig. 3b. In so doing the “0 cycle” curve is taken as the reference. It is seen that just after application of 5 basic cycles the shape of the loops, as well as their area and ε value tends to possess the same pattern as one corresponding to the reference cycles. The same measurements were performed at the distance of 630 μm behind the crack tip. These hysteresis loops have similar shape. It means that the application of the underload does not exert significant influence on them. In order to avoid overburdening of the manuscript we do not provide graphs for this case, however the computed value of ε max are presented in Fig. 4b. By calculating the values of the ε max parameter one can reveal quantitative difference of the obtained results. The plots of ε max vs the number of loading cycles are displayed in Fig. 4 for the measurements: (a) at the crack tip and (b) at the distance of 630 μm behind the tip. It is seen that two graphs demonstrates a pronounced sharp peak at “0 cycles” time step which suggests to be related to the influence of the underload cycle and formation of residual strains in the vicinity of the crack tip. As it comes from the Fig. 4b the near-tip stresses have a small impact on the crack edge opening displacement far away from the tip. The value of the latter at the peak is 20 % higher than that in the reference points for measurements conducted at the crack tip. At the same time at the distance of 630 μm it is just 3 % larger (that might be appointed to the noise level). Apparently, after the drop of the ε max value at the crack tip, it is followed by the ascending trend due to the smaller K op and ΔK th . 3. Results and discussion