General

It is generally accepted that in welded structures subjected to repeated external loads, microcracks may be initiated rather early in the fatigue life. Based on observations on the development of fatigue cracking, it has become common practice to consider fatigue life as consisting of three phases. These are initiation of a crack, propagation of the crack, and final fracture. Final fracture is simply the fracture under the last up-loading tensile load cycle and may be treated by assessment of unstable fractures, as presented in Chapter 16. However, the distinction between the first two phases is not very clear. The question arises of how to determine that the crack is so large that its growth can be properly defined in laboratory testing and in crack growth analysis. Fatigue crack propagation is understood here to mean the growth of cracks that are so large that the continuum mechanics approach can be applied.

The proportion of fatigue life of a structure that is spent in each of the two first phases depends on the material, the geometry of the detail being considered, and the loading. The initiation period for a fatigue crack in the base material without significant notches is relatively long in comparison with the propagation period. In contrast, at weld toes in structures with more severe stress concentrations, the formation of a dominant crack occurs relatively early in the fatigue life, and the propagation phase constitutes the major portion of the total life. Radencovic (1981) and test data on tubular joints reported by Pozzolini (1981) showed that 70–90% of the fatigue life of welded connections is related to fatigue crack growth. However, as indicated in the preceding paragraph, such numbers also depend on the definition of initiation and the type of connection. At start of fatigue crack growth after initiation it is assumed that a sharp crack tip has developed. The size of the crack at this stage is not so easy to define and in literature it ranges from below 0.1 mm up to 1 mm; see also Sections 3.1.3 and 4.7.4.

The basic mechanism of crack nucleation in the base material is cyclic slip and the extrusions and intrusions at the surface of the base material. At weld toes, however, initiation is more usually from defects at undercuts or from other imperfections in welded connections.