PSI - Issue 13

Yoshiki Mikami et al. / Procedia Structural Integrity 13 (2018) 1804–1810 Author name / Structural Integrity Procedia 00 (2018) 000–000

1805

2

1. Introduction Crack tip opening displacement (CTOD) tests are usually performed to ensure the structural integrity of high strength thick steels welds. However, weld residual stress inevitably exists in thick steel welds, and it affects the fracture toughness testing procedure and test results. For example, a fatigue precrack should be introduced in the specimen before testing as provided in testing standards (ISO 12135, BS 7448 Part 1, and WES 1108, for example). However, precracking is strongly affected by the distribution of weld residual stresses. A major effect of weld residual stress is that fatigue cracks do not initiate or propagate under a compressive residual stress field. Consequently, the preparation of precracked fracture toughness specimens itself becomes difficult. Even if a specimen is successfully precracked, the shape of the fatigue precrack front should satisfy the prescribed condition. In order to mitigate the effect of residual stresses, modification techniques such as local compression and reverse bending are adopted (ISO 15653, for example). Local compression is widely used and the required load to perform the local compression treatment can be estimated using Eq. (1): P LC = 1.4 B 2 R p0.2 , (1) where P LC is the required local compression load, B is the plate thickness, and R p0.2 is the proof stress of the steel. The application of the local compression technique is expected to be difficult owing to the increase in the thickness and strength of steels. On the other hand, Mikami et al. (2017) reported that the reverse bending technique is an effective method for residual stress modification with a considerably low loading capacity; the load required to perform reverse bending was lower than one-tenth that of local compression for the treatment of a plate with a thickness of 50 mm and a proof stress of 380 MPa. The residual stress distribution can be modified by the abovementioned treatment techniques; however, the resultant residual stress distribution depends on the treatment condition and cannot be completely zero. Therefore, it is important to track the variation in stress distribution through the steps of multipass welding and residual stress modification and to investigate the effect of residual stress distribution on the CTOD test result. Such an investigation needs the incorporation of residual stress distribution; however, fracture mechanics simulations have been rarely conducted considering the complex residual stress distributions that exist in actual welds. The authors developed a simulation method to model the entire process of welding, specimen machining, and fracture toughness testing. In this study, the variation in stress distribution throughout the processes of multipass welding, specimen machining, residual stress modification, and fracture toughness testing was simulated, and the effect of weld residual stress on P – V g (the load and notch mouth opening displacement, respectively) curve and the crack opening behavior in CTOD tests was investigated. 2. Fabrication of multipass welded joint and residual stress measurement EH40 (390 MPa-class in the yield stress) steel and the corresponding welding wire, which are used as materials for marine and offshore structures, were used to fabricate the multipass welded joint. Steel plates (50 mm thick × 250 mm wide × 500 mm long) were multipass butt welded by flux cored arc welding (FCAW). The geometry and dimensions of the groove is shown in Fig. 1(a). The welding conditions were a welding current of 240 A, an arc voltage of 25 V, a welding speed of 45 cm/min, preheat and interpass temperatures lower than 100 °C, and a shielding gas Ar+20%CO 2 . These welding conditions were determined based on the pre-qualification test prescribed in the American Petroleum Institute Recommended Practice 2Z (API RP 2Z). The cross-section of the multipass welded joint is shown in Fig. 1(b). Residual stress distribution through the thickness in the weld metal was measured using the strain gauge technique proposed by Takahashi et al. (1979). Strain gauges were attached on the joint as shown in Fig. 2(a); Two plates, T and L, were cut out from the joint and additional strain gauges were attached on the cut surface. The plates were cut into small pieces to measure the strain relieved, and then, the stress was calculated. The obtained residual stress distribution is presented in Fig. 2(b). The measurement result shows a typical residual stress distribution in multipass welded joints of thick plates. The tensile–compressive–tensile transverse residual stress ( σ y ) distribution through the thickness affects the fatigue precrack initiation and propagation from the machined notch root.