PSI - Issue 75

Kevin Mouradian et al. / Procedia Structural Integrity 75 (2025) 616–624 Kevin Mouradian et al. / Structural Integrity Procedia (2025)

619

4

A 3D scan was conducted on each specimen using the GOM ScanCobot system to verify the surface geometry of the welds. Furthermore, radiographic testing was conducted to examine internal weld imperfections. Annex B of ISO 5817 incorporates additional criteria for welds subject to fatigue loading, refining the acceptance limits for imperfections. It establishes quality levels B90 (FAT 90) and C63 (FAT 63) by adjusting permissible imperfection limits. For imperfections examined in this paper, the limits remain the same between quality levels C and C 63, as well as quality levels B and B 90. Notably, ISO 5817 does not define tolerances for angular misalignment and weld toe radius for static resistance. However, both imperfections were addressed in Annex B. 3.2. 3D scan test results The 3D scan enabled the precise measurement of several parameters (see Table 2), including the weld toe radius r , weld toe angle α , undercut height h und , linear misalignment h ali , angular misalignment α ali , and excess weld metal h w . The data were collected at intervals of 0.2 mm along the weld length. The weld toe angle on both sides of the weld was measured and the results are presented in Fig. 2 as a scatter plot. In all specimens, the plate surface served as a reference for weld toe angle measurements. The specimens designated as BW_TYPE exhibited toe angles ranging from 148.1° to 180°, while those designated as BW_UND ranged from 143.9° to 180°. The mean toe angle measured on each specimen was found to meet the minimum requirement of 150° specified for quality level B (B90). However, small zones in some welds exhibited angles below 150°, corresponding to quality level C. a b

Fig. 2. Weld toe angle distribution and quality levels in (a) BW_TYPE; (b) BW_UND

The 3D scan enables the identification of undercuts on both the left and right sides of the weld. Fig. 3 illustrates the height of the larger undercut measured across both sides along the full weld length. Undercuts were also identified within the BW_TYPE specimens, even though it was intended to comply with quality level B (see Fig. 3a). The undercuts were discontinuous, extending over 10 mm to 17 mm of weld length, with the h und reaching up to 1.1 mm. The BW_UND specimens exhibited continuous undercuts along the entire weld length, as illustrated in Fig. 3b, with maximum undercut height ranging from 1.6 mm to 3.0 mm. The larger undercut was observed in specimen UND_1, while UND_5 and UND_6 exhibited the lowest values. A notable observation was the variation in undercut height along the weld length. In some locations, the undercut complied with quality level B limits (0.5 mm), while in others, it significantly exceeded the thresholds defined for quality level D (1 mm). This variation raises a critical question whether the maximum measured imperfection or the average value along the weld should govern the determination of the weld quality level. In this study, the authors based their evaluation on the mean values of undercut.