PSI - Issue 13

Taiko Aikawa et al. / Procedia Structural Integrity 13 (2018) 104–109 Aikawa/ Structural Integrity Procedia 00 (2018) 000 – 000 fracture near the plate surfaces might be replaced by the increase in crack arrest toughness. Hence, in the present model, is assumed to have a distribution in the thickness direction, as schematically shown in Fig.4. Note that this curve depends on the magnitude of . Because the uncracked ligament carries crack closure stress, stress intensity factor is not constant along the crack front, but it decreases near the plate surfaces, see Fig.4. Using the equation of stress intensity factor for a crack subjected to the crack closure stress was applied for calculating the distribution of stress intensity factor (Broberg, 1999). The present model is quasi-static and does not include dynamic (inertial) effect associated with crack propagation. 107 4

4. Calculations 4.1. Determination of parameters

Numerical simulations were conducted to determine the values of the parameters contained in the model. Brittle fracture surface of temperature-gradient crack arrest test of 30mm thick steel plate having yield strength 355MPa was subjected to the analysis. Specimen width was 500mm and temperature distribution in the specimen width direction was applied with temperature gradient of 0.3 ℃ /mm. Applied stress was 300MPa. Arrest crack length was 302mm. Figure 5(a) shows fracture surface. Note that only the last half of the overall fracture surface is shown. A 3-D laser measurement system was applied to capture 3-D fracture surface morphology of the fracture surface and the regions of brittle fracture surface and shear-lips were identified, as shown in Fig.5(b).

Fig.5 Fracture surface of temperature-gradient crack arrest test, (a) photograph, (b) 3-D laser scanned image. Figure 6 shows results of numerical simulations of the above test, where the shape of the solid line in Fig.4 and the value of were changed. Note that the parameter f represents the steepness of the solid line. In Fig.6(b), the effect of the loss of plane strain was too weak so that the shear-lip depth was too shallow, compare with Fig.5. But in Fig.6(a), the depth and shape of the shear-lips were reproduced well, in which case = 1.5 and = 0.3 .

It should be noted in Fig.6(a) that the development of the shear-lips with increasing crack length was well reproduced. In addition, the fracture surface irregularity, including the chevron markings, was well reproduced. It should also be mentioned that the arrest crack front shape is well reproduced, i.e. , the crack propagated longer at mid- Fig.6 Calculated fracture surface, (a) = 2.0 , = 0.3 , (b) = 1.5 , = 0.3 .