PSI - Issue 13

Fuminori Yanagimoto et al. / Procedia Structural Integrity 13 (2018) 2095–2100 Author name / Structural Integrity Procedia 00 (2018) 000–000

2096

2

According to a few studies on structural crack arrest design, crack front shape in the joint has to be considered to explain crack arrest phenomena in the joint (Handa et al., 2015, 2014). Namely, it is assumed by the previous studies that the crack front becomes to reduce stress intensity factor at the crack tip in the joint structure. However, although the crack front shape behaviour is significant to develop structural crack arrest design as pointed above, the crack front shape in such 3D joint structures has not been observed and the effect of crack front shape behaviour has not been proven yet. This is because steels are not transparent. Strain gauges and timing wires to detect rapid crack propagation employed by previous studies (Prabel et al., 2008; Yanagimoto et al., 2018) can only measure 2D crack behaviours and are not able to measure 3D rapid crack propagation behaviour. In this study, considering that brittle polymer is often employed to study fast crack propagation and arrest behaviours due to its simplicity (Dally et al., 1985; Nishioka et al., 2010), we carried out direct optical observation of fast crack propagation and arrest behaviour including crack front shapes in 3D joint structures composed of transparent polymer, PMMA, which is often used for dynamic fracture mechanics researches (Ayatollahi et al., 2016; Grégoire et al., 2009). This experimental approach is carried out to investigate 3D dynamic fracture mechanics in complicated 3D structures for the first time although Nishioka et al. referred 3D dynamic crack propagation in simple plates (Nishioka et al., 1985). 2. Experimental preparation 2.1. Materials Polymethyl methacrlate (PMMA) was employed in this study to simulate brittle crack propagation and arrest behaviors. PMMA is brittle elastic polymer. Comoglass, which was made by Kuraray, was usually employed as PMMA, but EX001, made by Mitsubishi Chemical, was used when the thickness was 5 mm and 10 mm. Mechanical properties of them are shown in Table 1. Their properties are almost same although their product names are different. Usually, crack arrestability of PMMA is written by stress intensity factor. �� , which is crack arrest toughness when the crack velocity is zero, is regarded as a material constant (Grégoire et al., 2009). �� of the employed PMMA was identified as 1.16 MPa√m by DCB plate crack arrest test. This value is almost same as ones obtained in the past study (Grégoire et al., 2009; Yue et al., 2017). Table 1 Mechanical properties of employed PMMA Name t [mm] TS [MPa] Young's modulus [MPa] Fracture strain [%] 2.2. Specimen and set-up of high speed camera The specimen geometry is shown in Fig. 1. This imitated structural crack arrest design used for large container ships (Handa et al., 2015). Although it is desirable to provide tensile loading to the specimen because the tensile loading is expected to the actual structure, the DCB-type specimen was employed to enable the high speed camera to observe 3D crack front shape in the joint structures. Thus, the stress intensity factor decreases along crack propagation. A web was connected to a flange by solvent bonding. Surfaces of the structures were polished to improve their transparency. A wedge was used to provide force displacement to the pins to initiate a crack. w , width of the initial notch was set to 1.2 or 1.6 mm. The curvature of the initial notch was 0.5 . The thickness of web plate, � , was changed from 5 to 13 mm to investigate the interaction between crack front and the structure. Because the duration between crack initiation and arrest is extremely short, the specimen is under the fixed displacement condition. In order to observe fast crack propagation and arrest behavior in PMMA, a high speed camera, SHIMADZU HPV 2, was employed. The frame rate was set to 250,000 per second ( 4μs per frame) and it can observe 100 frames (400 μs ). The camera observation was initiated by cutting of a crack gauges glued on the specimen surface. Charpy absorbed energy [ kJ m � ⁄ ] 17 density [ kg m � ⁄ ] 1170 Rayleigh wave velocity [ m⁄s ] 954 EX001 5, 10 74 73 3200 3200 4.5 5.0 Comoglass 13 17 1170 954