Crack Propagation in Weldments

Abstract

The basic topic of the present study is the determination of the fatigue strength by fracture mechanics. Common fracture mechanics-based fatigue considerations, e.g., within the frame of damage tolerance considerations, are usually limited to residual lifetime determination of so-called long cracks which are often defined in conjunction with nondestructive testing. The extension of this concept to the total lifetime, as in the S-N curve approach, requires an adequate description of so-called short crack propagation which cannot be based on the DK concept and must consider the crack closure phenomenon as well as its gradual buildup at the short crack stage.

Further, it has to provide a meaningful definition of the initial crack dimensions needed in fracture mechanics and a solution for the multiple crack problems at stress levels higher than the fatigue limit as this is specific for configurations such as weldments. A fundamental discussion of these aspects beyond its application to welds is provided by Zerbst et al.

The idea to apply fracture mechanics to the determination of the fatigue strength of weldments is anything but new. First attempts date to around 1970 (Maddox 1970) who wrote in 1974: “It is now widely recognized that flaws will inevitably exist in welded structures and the old idea of removing all detectable defects must be replaced by the ‘fitness for purpose’ design philosophy. This makes it necessary to define reliable methods of assessing the significance of flaws, particularly in the context of fatigue, …. The most promising approach to this problem lies in the use of the fracture mechanics based description of fatigue crack propagation.” (Maddox 1974). At that time, a major problem was to develop appropriate K-factor solutions. Ever since effort was spent to further develop the approach. No detailed overview shall be given here; see, however, Radaj et al. (2006), Chap. 6, for a review. Note that even today, fracture mechanics besides some promising applications (see, e.g., Nykänen 2005, repre- sentative for many other studies) is still not widely used as a tool for fatigue strength determination of weldments. One reason for this is certainly that the approaches usually fail to meet the above-formulated requirements. They are based on linear elastic fracture mechanics even for short cracks and on fixed or (from S-N curves)back-calculated initial crack sizes, they usually consider one crack only, neglect the variation in the local weld geometry along the weld toe, etc.

Fatigue cracks are initiated at the smooth surface in plain specimens (i.e., at surface extrusions and intrusions caused by persistent slip bands) or—more frequently—at geometrical discontinuities (which act as geometric stress concentrators) and/or material defects such as non-metallic inclusions (which cause strain concentration zones due to their different stiffness compared to the matrix material and sometimes also chemical mis-match), see Zerbst et al. (2018b). Polak (2003) notes with respect to the crack nucleation stage: “Numerous studies have shown that in the majority of materials and under normal loading conditions, the period of crack initiation in smooth specimens without defects amounts to less than 5–20% of the fatigue life. In materials containing defects, the fraction of life spent in crack initiation is even lower. The major part of the life is spent in the growth of cracks, namely in the growth of short cracks.” That this statement usually is also true with respect to welds has been confirmed, e.g., by Verreman and Nie (1996) for manual fillet welds of structural steel.

Keywords: Crack Propagation, Weldments, fatigue strength, short crack propagation, Crack Initiation