Since the determination of S-N curves in accordance to standard DIN 50100:2016-12 is very time-consuming and cost-intensive as a result of the large number of required specimens, methods for an accelerated lifetime prediction have been developed recently. In order to increase the quality of these methods, statistical approaches need to be taken into account. Therefore, it becomes necessary that the considered distribution function describes the statistical nature of the phenomenon, which has to be analyzed. As there are variations regarding symmetrical and asymmetrical distributions in fatique, depending on the material characteristics (e. g. roughness, residual stresses, grain orientation), suitable statistical tests must be applicated for a reliable assessment. The aim of the presented research is to ensure a reliable consideration of the material-dependent scatter and a comparability of accelerated and conventionally determined scatter band.

The already developed lifetime prediction method StressLife is used and combined with statistical approaches. To obtain experimental data, fatique tests concerning SAE 1045 are carried out, including non-destructive measurement methods (for example the following methods can be used: optical strain measurement, electrical resistance measurement, thermography). Apart from that, statistical evaluations are integrated in several steps. First, a normal distribution test is performed to determine whether the data are parametric or non-parametric. This is done using the Shapiro-Wilk, Kolmogorov-Smirnoff and Anderson-Darling test. Depending on the result of the normal derivation, the suitable significance test is carried out in the next step. Furthermore, the Weibull distribution are also checked. At least, the distribution function is determined and the survivor function can be derived from it.