A new statistical software for the estimation of P-S-N curves in presence of defects: statistical models and experimental validation

It is well-known in the literature that fatigue failures originate from the weakest element within the component loaded volume. In particular, for metallic materials the fatigue crack typically nucleates from the most critical surface defect in the High-Cycle-Fatigue (HCF) region; whereas, it generally nucleates from the most critical internal defect in the Very-High-Cycle-Fatigue (VHCF) region, at stress amplitudes below the so-called 'transition stress amplitude'. Therefore, regardless of the fatigue region, the P-S-N curves must necessarily take into account the random distribution of the critical defect size. However, in the literature, there are few statistical models that can model the dependency between the fatigue life and the defect size or that can take into account the presence of different failure mechanisms (e.g., surface crack nucleation in HCF and internal crack nucleation in VHCF). In the present paper, a new software for the estimation of the P-S-N curves is proposed. The statistical P-S-N models recently proposed by the authors and implemented in the software are described in the paper. The procedure for the parameter estimation is also explained in detail and the software is finally validated with experimental datasets collected by the authors or available in the literature.