Font Size: 

The current philosophy to apply the safe life approach to the design of the railway axles does not
guarantee an infinite life due to the arising, too soon and especially, of completely random local damages that
can grow under mechanical and ambient service loading up to compromise the strength of the component. The
effects sometimes catastrophic of such event push for hold the attention in facing the problems related to the
design of these particular components by the Agencies and the manufactures involved. Among all, the most
important aspects are the analysis of the real reasons of the failure, on one side, and the definition of rules and
practices related to the design and maintenance of the components, on the other side. Indeed, it is not necessary
to invent new theories or techniques, but to improve or enhance those already knew, to contribute to improve
the quality of the component and the safety of its use. In this scenario, it is justified any attempt to improve the
modelling of the fatigue crack growth phenomenon that can propagate from a geometric singularity originated
on the axle. A better understanding of the phenomenon can contribute to significantly guide the design choices
toward an extended use of the Damage Tolerance philosophy, which will completely replace the actual Safe Life
approach once adequate levels of both reliability and NDT costs, used to periodically verify the structural
integrity of the component, will be achieved. A lot of technical papers deal with this problem, each of them to
study (or examine in deep) the effects produced by one or more factors on the fatigue crack growth, namely:
loads, constraints, environment, geometry of the initial defect, numerical modelling technique and experimental
method on standard specimens or full scale components. Focusing the attention on the building up of the
propagation model, it can be stated that, over the years, numerical procedures have became more and more
complete and complex in an attempt to reproduce the phenomenon in a more realistic way. However, there is
still a strong need of refinement of the current modelling techniques. In almost all published papers produced
until now, the growth of the crack has been studied assuming only a planar propagation, on the original plane
used to define the defect, and an elliptical geometry of the front, which keeps its shape during the whole
propagation process. However, in reality, the crack surface it is not always planar and the crack front evolves
toward geometry much more complex, even if it can be assimilated, with a good approximation, to an ellipse
during the initial part of the propagation phase.