Abstract
Many fatigue damages in welded built-up structures are often found in stress concentration sites, and fatigue cracks are initiated from stress concentration sites even though no initial defect exists in the region. Hence, it is required for quantitative assessments of fatigue life for the estimation of the continuous fatigue crack growth curves from initial defect-free sites in stress concentration sites. Fatigue cracks in welded built-up structures are usually initiated as plural surface cracks at a stress concentration site such as a fillet weld toe. These cracks coalesce as they advance, and finally single large surface crack appears. It is observed that the depth of the fatigue surface crack just after the completion of the coalescences is usually very small such as 0.5mm or so. By considering such characteristics of fatigue cracks initiated from weld toe, Toyosada et al. proposed the representative surface crack concept to perform the fatigue life estimation of the structures based on the advanced fracture mechanics approaches. The rule of shape evolution for the representative surface crack must be given in order to perform the fatigue life estimation. Then, Toyosada et al. proposed the empirical formula, which based on the observation of surface cracks growth histories in the out of plane gusset joint with boxing fillet weld. It is, however, considered that their formula cannot be applied in case that wide fatigue cracks initiation sites. Authors propose the conventional estimation formula of the shape evolution of a representative surface crack, which has broad utility, in this study and conduct the validity by comparing the measurements of shape evolution.
