For many components fatigue strength is the limiting factor. Therefore, proper understanding of how different factors impact fatigue strength is important to make use of the full potential of a material. Too low strength and the part will fail, and an overly conservative design will be bad from both a cost and sustainability point of view.

With models that can describe how the structure of the material affects the strength under different conditions, it’s possible not only to make the best use of materials, but also to find new ways to improve the strength.

In many cases there is a direct correlation between fatigue strength and defects, such as inclusions, in the material. In PM steels fatigue crack typically initiates at the pores, which act as local stress concentrations. A fatigue model for PM steels should therefore include the effect of pores. Often, this is done indirectly by taking strength as a function of density. However, such an approach does not give any insight into the actual mechanisms acting and will not help describe new phenomena.

This paper will describe how a fatigue model for PM steels can be developed, where a statistical model of the porosity is linked to the fatigue strength through a fracture mechanics approach. Through a couple of examples it will also be demonstrated how the model can be used to take various effects into account. For instance, both density dependence and the notch effect are automatically considered, but it is also possible to include for instance different pore sizes generated by sieving out finer powder fractions. The model presented here focuses mainly on heat treated materials with a hard-micro structure.

Developing the model

The basic assumption of the model is that the fatigue crack will initiate at the largest pore in the highly stressed volume. This corresponds to saying that the material will fail at the weakest link. The concept of highly stressed volume corresponds to saying that a large pore in an unstressed area will not initiate. Often the highly stressed volume is taken as the part of the material with 90% or more of the peak stress in the component. The 90% value is somewhat arbitrary but has proven to work well in practice and will also be used here.

To find the largest pore it is necessary to build a statistical model of the porosity. Here extreme value statistics is a powerful tool since the largest pore rather than the average pore should be determined.

Log in to your free profile to access this article.

Already a Materials Today member?

Log in to your Materials Today account to access this feature.