On the estimation of the fatigue cycle distribution from spectral density data

Abstract

This paper deals with the fatigue life prediction of components and structures subjected to random fatigue, i.e. to cyclic loading whose amplitude varies in an essentially random manner. In particular, this study concentrates on the general problem of directly relating fatigue cycle distribution to the power spectral density (PSD) by means of closed-form expressions that avoid expensive digital simulations of the stress process. At present, all the methods proposed to achieve this objective are based on the use of a single parameter of the PSD. In this work, by numerical simulations and theoretical considerations, it is shown that the statistical distribution of fatigue cycles depends on four parameters of the PSD and the methods proposed in the literature provide reliable results only in particular cases. The proposed approach lays the foundation for a more accurate evaluation of the fatigue cycle distribution. As a first result, a closed-form solution for the case of broad-band processes with narrow-band time derivative has been obtained in this paper.