The significance of fatigue crack initiation for predictions of the fatigue limit of specimens and structures

• Differences between micro-crack initiation and macro-crack propagation are highlighted.
• Current predictions on the fatigue limit of notched elements are entirely empirical.
• FEM calculations on stress distributions in a structure are a powerful tool for design purposes.
• Predictions on fatigue of welded joints is a most complex problem.
• The current rules for obtaining S–N curves of welded joints do not agree with the present knowledge.

The fatigue life of specimens and structures covers two periods: a crack initiation period and a crack growth period. Micro-crack nucleation and initial micro-crack growth are a surface phenomenon controlled by the local stress cycles at the material surface. The subsequent macro-crack growth is depending on the fatigue crack growth resistance of the material as a bulk property. The fatigue behaviour in both periods is qualitatively reasonably well understood. However, the quantitative analysis is problematic. Moreover the number of variables which can effect the fatigue behaviour of specimens and structures is large. The paper is focussed on realistic understanding of the prediction problem, especially on the prediction of the fatigue limit of notched specimens and structures. The effect of a salt water environment on the fatigue limit is discussed. As a special topic comments are presented on the notch effect of welded joints. Short comings of the so-called effective notch concept are indicated. Comments on the design recommendations of the International Institute of Welding are presented. The significance of realistic experiments and a profound FE-analysis are emphasized.