PSI - Issue 42

Vitor S. Barbosa et al. / Procedia Structural Integrity 42 (2022) 1177–1184 V. S. Barbosa and C. Ruggieri / Structural Integrity Procedia 00 (2019) 000–000

1181

5

conducted on three-point single edge-notched specimens, commonly designated as SE(B) configurations, in the T L orientation extracted from the mid-thickness of the plate. The SE(B) specimens have conventional, plane-sided geometry with thickness, B = 25 . 4 mm (1T configuration), width, W = 50 . 8 mm, loading span, S = 203 . 2 mm, and a nominal crack length ( a ) to width ( W ) ratio of a / W = 0 . 5. An important first step in applying the MC procedure involves the selection of an adequate test temperature from which the measured fracture toughness values can be used to determine T 0 . Following ASTM E1921 (American Society for Testing and Materials, 2019), a convenient estimate of the indexing temperature, ˜ T 0 , for ferritic steels is given by ˜ T 0 = T 28J − 18 o C or ˜ T 0 = T 41J − 24 o C, which yields T 0 − 28J = − 20 . 8 o C and T 0 − 41J = 66 o C for the martensitic steel under consideration. If we tentatively accept the T 0 − 28J estimate as being close to correct, then we can set T = − 20 o C as the initial test temperature. However, the load vs. CMOD curves for three specimens tested at T = − 20 o C, revealed an essentially linear elastic brittle behavior with almost no plastic work under the load displacement curve - see Barbosa et al. (in press) with details of the experimentally measured load vs. CMOD curves. Because the MC procedure requires elastic-plastic behavior, it is thus necessary to increase the test temperature, such as the fracture tests were performed at the temperatures T = − 20 o C, − 10 o C, 0 o C, 20 o C, 50 o C, 60 o C and 70 o C. In view of elastic-plastic behavior, only the fracture toughness tests conducted at T = 50, 60 and 70 o C have shown this behavior until load instability. With those factors taken into account, only the fracture toughness measured at T = 50, 60 and 70 o C will be used in the present study. Figure 2 displays the cumulative probability distribution of the fracture toughness values in terms of K Jc -values for the tested specimens at T = 50, 60 and 70 o C. The solid symbols represent the experimental fracture toughness data for the specimens. The curves displayed in these plots describe the 3P-Weibull distribution for K Jc -values given by previous Eq. (1) with a fixed value α = 4 to describe the scatter in test data as adopted by ASTM E1921 (American Society for Testing and Materials, 2019). Table 2 provides the ML estimates of the characteristic toughness, ˆ K 0 , for all tested crack configurations derived from Eq. (3), including the 90 % confidence bounds given by Thoman et al. (1969) (see also Mann et al. (1974)).

Fig. 2. Three-parameter Weibull distribution of experimentally measured K Jc -values with varying test temperatures at T = 50, 60 and 70 o C.

The key feature of these results is that there is a marked e ff ect of temperature on fracture toughness as the char acteristic toughness, K 0 , increases rather distinctly with increased test temperature. Unfortunately, the experimental results are not particularly well-fit by Eq. (1) with a fixed value α = 4. While we have not investigated the source of such behavior, we argue that the predominantly martensitic microstructure associated with potential changes in the local cleavage micromechanism compared to conventional ferritic steels (see discussion in the review article of Hahn (1984)) may be the cause of the reduced scatter in the experimental data - recall that the Weibull modulus charac terizes the scatter in test data. However, as already noted by Ruggieri et al. (2015), ensuring a high fitting quality of the three-parameter Weibull distribution to the experimental data is not a requisite feature for strict applications of