Review of the fatigue strength of welded joints based on the notch stress intensity factor and SED approaches

• Published N-SIF-based S–N curves are modified to consider misalignment effects.
• For the SED approach, a larger control radius is proposed to consider misalignment.
• Own tests with relatively simple V- and U-notched specimens are presented.
• The crack propagation phase determines the fatigue life of U-notched specimens.
• Hence, single-point fatigue parameters can result in over- or underestimated lives.

Several approaches exist for the fatigue strength assessment of welded joints. In addition to the traditional nominal stress approach, various approaches were developed using a local stress as fatigue parameter. In recent times, the N-SIF based approaches using the notch stress intensity at the weld toe or root have been developed. Based on this, the more practical strain energy density (SED) and the Peak Stress approaches were proposed. This paper reviews the proposed design S–N curves of the N-SIF and SED approaches questioning in particular the consideration of misalignment effects, which should be included on the load side of local approaches in order to consider them individually in different types of welded joints. A re-analysis of fatigue tests evaluated for the effective notch stress approach leads to slight changes of the design S–N curves and of the radius of the control volume used for averaging the SED at the notches. Further, on purpose fatigue tests of artificially notched specimens show that the fatigue assessment using a single-point fatigue parameter might be problematic because the crack propagation phase, being part of the fatigue life, is strongly affected by the stress distribution along the crack path that may vary considerably between different geometries and loading cases.