PSI - Issue 13

Yuki Nishizono et al. / Procedia Structural Integrity 13 (2018) 1817–1827 —‹ ‹•Š‹œ‘‘ Ȁ –”—…–—”ƒŽ –‡‰”‹–› ”‘…‡†‹ƒ ͲͲ ሺʹͲͳͺሻ ͲͲͲ – ͲͲͲ ͵ range, the lack of correlation with the results evaluated by the wide plate tensile test, and the need to accumulate large amount of data for each steel. Thus, there is still a substantial need for an improvement of simplified evaluation method based on fracture mechanics. Recently, the local fracture stress criterion that the local stress in the vicinity of crack tip is constant has been demonstrated for the dynamic crack propagation (Machida et al., 1992), it is suggested that the explanation of crack propagation/arrest behavior in the conventional evaluation method which assumes that crack driving force is increasing with crack propagation in tensile specimen, may be different from the actual phenomenon. Actually, based on the local fracture stress criterion, it seems to be possible to acquire the local fracture stress of each steel by precisely simulating some kind of small-scale test by FEA and apply it to the safety evaluation of steel structural components. However, the authors think that such an approach will not immediately lead to industrial interests as compared with the conventional approach which has shown effectiveness for a long time. In this study, considering the simplicity and connectivity with abundant existing data, the authors applied the conventional explanation that the magnitude relation between crack driving force and crack propagation resistance determines brittle crack propagation/arrest. This is because the final target of this research is to estimate arrest toughness ca evaluated by the wide plate tensile test by conducting small-scale bending test. In order to ensure the affinity with the conventional approach, the authors applied a dynamic stress intensity factor calculated by a devised and simplified dynamic 3D elasto-plastic FEA as a parameter characterizing crack driving force. In other words, the authors recognize the devised FEA is slightly different from the real situation around a running crack. Test steels used in this study were two types of steel plates which was produced by normalizing. Steel A was used in the preliminary experiments and steel B was used in the main experiments. The chemical composition and basic mechanical properties of these steels are listed in Table 1 and Table 2, respectively. The original thicknesses of these steels are 30 mm and 28 mm, respectively. In order to acquire the constitutive equations, several tensile tests under different temperatures and strain rates were performed (Tonsho et al., 2015). Configurations of stress-strain curves and rate dependencies were approximated by Swift equation (Eq. 2) and Cowper-Symonds equation (Eq. 3), respectively. Table 3 and Table 4 show the configuration results. The authors applied these material properties to dynamic 3D elasto-plastic FEA. Table 1. Chemical composition of the two steels. Steel C Si Mn P S A 0.140 0.410 1.45 0.017 0.003 B 0.135 0.340 1.40 0.009 0.002 1819 2. Test steels

EL [ % ] vTrs [ ° ] 27 -40

Table 2. Mechanical properties of the two steels.

YS [MPa]

TS [MPa]

Steel

A B

416 368

541 503

45

-104

Table 3. Configuration results of Swift equation. Temperature [ ° ] [MPa] -140 649 0.02 0.024

Steel

0.172 0.236 0.263 0.279 0.178 0.182 0.146

-100

501 433 394 595 535 476

0.02 0.02 0.02

0.019 0.013 0.015 0.0217 0.0126 0.0156

A

-60 -20

-100

0.023 0.019 0.013

B

-50

20

Table 4. Configuration results of Cowper-Symonds equation. Steel C P A 5.04 × 10 4 5.14 B 5.50 × 10 4 6.25