PSI - Issue 13

Fuminori Yanagimoto et al. / Procedia Structural Integrity 13 (2018) 116–122 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

117

2

These requirements are based on “Long brittle crack problem”. This problem means that brittle crack can be arrested even if the stress intensity factor (SIF) is higher than crack arrestability when the SIF is extremely high. The reason why the crack is arrested under extremely high SIF has been unclear and remain one of the unsolved problems in brittle crack propagation and arrest behaviour in steels (Machida and Yoshinari, 1986). Since “Long brittle crack problem” plays an important role in the requirement , it is essential to investigate the behaviour of the brittle crack with high SIF. According to a numerical model based on the local fracture stress criterion to simulate brittle crack propagation and arrest behaviours proposed by Shibanuma et al. (Shibanuma et al., 2016a), a rise of SIFs leads to growth of unbroken shear lips to decrease effective SIFs and the crack can be arrested even when the SIF is extremely high. Considering that the calculation of this model could explain the crack arrest experimental results including some high SIF conditions, the concept that the unbroken shear lip growth due to rise of the SIFs play an important role in brittle crack propagation and arrest behaviours is reasonable. However, the details of crack behaviours under high SIF has not been clear, and the condition of shear lip formation in the model was based on a bold assumption. Its validity was not investigated especially when the crack was much long. In fact, although the model calculation result indicates that there are “upper curve ” of the relationship in addition to “lower curve” between crack arrestability and temperature as shown in Fig. 1, there are only a few experimental data to support such prediction. Accordingly, it is valuable to examine brittle crack propagation and arrest behaviours in extremely high SIF condition using wide specimens. In order to examine the crack behaviour under high SIF, we carried out experiments where the SIF was always larger than the lower curve of crack arrestability obtained in usual crack arrest tests. Fig. 1 shows a schematic illustration of SIF transition in the experiments. Thus, the SIF in the wide specimen was planned to keep higher than lower curve of the model prediction curve on the crack arrestability and it was investigated whether the crack was arrested when the SIF reached the upper curve. Therefore, the model calculation was carried out prior to the experiments. Although in fact, a so called duplex specimen has been employed to evaluate brittle crack propagation and arrest behaviour in high SIF, such specimen includes welding and discontinuity of materials because the specimen is composed of a brittle crack running plate and tough test plate. Because these factors influences the crack behaviours, it is hard to interpret the experimental results in the duplex specimens. Here, this study employed homogeneous wide specimens to investigate brittle crack propagation and arrest behaviours under high SIF conditions. At first, we implemented temperature gradient crack arrest tests to obtain a crack arrest toughness of the steel employed. Second, a calculation of the model mentioned above was conducted based on the crack arrest toughness to determine experimental conditions. And then, the wide specimen experiments were conducted.

500

400

Upper curve

0 Crack arrest toughness [MPa m] 100 200 300 3.6

Temp. Grad. Arrest Test Model prediction SIF transition in wide specimen

Lower curve

3.8 1000/ [ −1 ]

3.7

3.9

4.0

Fig. 1 Schematic illustration of model predictions and SIF transition in wide specimen

2. Preparation of experiments 2.1. Steel characteristics The ferrite-pearlite steel with 30 mm thickness was employed in this study. The chemical composition is shown in Table 1. Mechanical properties of the steel is shown in Table 2. The steel was a ferrite-pearlite steel.