Fatigue can be defined as the cracking of a material or component subjected to cyclic loads. The main problem is that the load required to initiate and propagate a crack is usually low. Consequently, the cracks grow very slowly at the beginning and they remain undetected most of the time. Once they grow larger, crack growth is accelerated and catastrophic failure may occur almost without notice.

The fatigue life of a component can be divided into two stages: crack initiation and crack propagation.

Crack initiation

Typically, cracks are initiated at the surface of a component which is subjected to cyclic loading. Initiation sites are usually associated to stress concentrators, such as dents, pits, holes…, i.e. surface defects in general. In addition, cracks may initiate in persistent slip bands due to fatigue cycling. Whatever the mechanism responsible for crack initiation, this is a random process. It is very difficult to predict where the crack will start because even minute defects can be very damaging. Most of the fatigue life of a component is spent in this stage. Sometimes, even 90% of the fatigue life is dominated by the initiation stage.

Crack propagation

This stage consists of the growth of a pre-existing crack in a component subjected to cyclic loads. Sometimes it is difficult to establish the border between both stages, i.e. when initiation ends and propagation begins. Usually, when a "long" crack is formed, it is considered that initiation has ended. However, this term is not defined in a unique way. For a metallurgist a "long" crack might be one larger than the typical micro-structural features of the material (i.e. grain size). On the other hand, an engineer might think that a crack should have at least 1 mm to be considered "long".

Crack propagation can be studied by applying Fracture Mechanics techniques. In fact, as the plastic zone size at the crack tip is usually very small in high-cycle fatigue, linear elastic fracture mechanics is used to characterize crack propagation, by means of the Paris law. One of the advantages is that standardized testing procedures are available for constant amplitude loading. As pointed out previously, the proportion of fatigue life spent in this particular stage is in general much less than in the initiation stage.

Shot peening is employed with the aim of improving the fatigue behaviour of components. It is thought that such improvement is due to compressive residual stresses and surface hardening. Both effects combine to hinder crack initiation. For example, compressive residual stresses extend in the subsurface region (typically up to 400-500 micron) and they reduce the local stress range applied to the component at the surface layer. Consequently, crack initiation will be delayed. Regarding surface hardening, if the surface is harder, crack initiation might occur in the subsurface region, but there, compressive stress will act as a wall.

The benefits of shot peening in the propagation stage are not so clear. For a "long" crack, stress is null at the crack flanks and there is a high stress concentration at the crack tip. As a consequence, the existing residual stress field is relaxed and stress redistribution takes place as the crack grows through the component.

For questions contact jesus@mfn.li