ABSTRACT
The question how to correctly evaluate fatigue test data in order to establish the reference value ΔσC of the characteristic fatigue strength of steel structures according to the safety requirements of Eurocode 3 still leads to discussions. This paper is supposed to give an overview of the state-of-the-art on the statistical analysis of fatigue test data. Firstly, the classical methods to set up different types of two-sided statistical intervals such as confidence intervals, prediction intervals and tolerance intervals and their one-sided bounds are addressed. The different meanings of these intervals and bounds are pointed out. Secondly, the statistical method pursued by Eurocode 0 is described which partly differs from the classical approach. Finally, the formal requirements of Eurocode 0 and Eurocode 3 are summarized, and it is highlighted in how far the aforementioned methods meet these requirements. To illustrate the application of the aforementioned methods, a particular test data set is assessed with forced slope of the S-N curve in a worked example. The reference values ΔσC obtained by the different methods for the test data set are compared.
