Design against fatigue failures: Lower bound P-S-N curves estimation and influence of runout data

In the present work, three methodologies for the estimation of the fatigue design P-S-N curves are compared. The approximate Owen with the staircase method, a method based on the likelihood ratio confidence interval, and an original method based on the bootstrap approach are considered. Three experimental datasets are employed for the experimental validation, showing that the three methods permit to have a reliable estimation of the design curves and that the number of data and, accordingly, the test duration can be significantly reduced with the proper testing strategy.