PSI - Issue 57

Gustav Hultgren et al. / Procedia Structural Integrity 57 (2024) 428–436 Hultgren & Barsoum/ Structural Integrity Procedia 00 (2023) 000 – 000

431

4

1

equ     = 

   

0 B dB



sec 98 where the sectional stress is the integration of the Signed von Mises stress criterion over that specific cross-section, 99 1 B    , (5)

0 L dL

sec    = 

  



SvM 100 The reference area is in Eq.(5) and Eq.(6) divided into a reference weld length component 101 profile length component 0 L 14.7mm = , of which their combined product becomes 0 A . 102    .

(6)

0 B 60mm = and a reference

3.2. Decreased scattering 103 The experimental fatigue test results in Fig. 2(a) show the high scattering for the nominal stress range with a 104 standard deviation in logC of 0.36. The high scattering is an anticipated result of the significant variation in weld 105 geometry between the specimens, which consequently lowers the fatigue strength as most specimens are below the 106 FAT 80 curve. Taking the true weld geometry into consideration through the numerical simulations and fitting the 107 Weakest-link method lowers the scattering notably, as seen in Fig. 2(b), with the standard deviation in logC reduced 108 by more than half to 0.15. The reduction implies that most of the variation in the fatigue test results was due to 109 geometric variations in the welds. 110 3.3. Weld geometry 111 The Winteria® measurement system enables a qualitative assessment of weld geometry by quantifying various 112 geometrical definitions, including the weld toe radius, weld toe angle, throat thickness, leg length, and undercut. Fig. 113 3 displays the kernel distributions of the measured weld geometry and the ratio between the sectional and nominal 114 stress for all the examined specimens. The samples are arranged in descending order from the least geometry-critical 115 specimen at the top to the most geometry-critical specimen at the bottom. 116 The top six rows in the figure exhibit a common characteristic of smoother weld toe transition radii on average. 117 These specimens also tend to have larger throat thicknesses and leg lengths, with minimal or no undercut, except for 118 the top two specimens, where small undercuts are observed. Åstrand et al. (2016) demonstrated that smooth undercuts, 119 which facilitate larger weld toe radii, mitigate the geometrical notch effect. The higher heat input in the peripheral 120 region of the molten zone responsible for the undercut also reduces the likelihood of cold-laps formation, commonly 121

Figure 2. SN curves for (a) the nominal stress range and (b) the equivalent stress range.