The probability of HCF – Surface and sub-surface models

•New HCF model based on highly loaded regions has been proposed.•The new model is based on sub-surface region below surface.•The new model has been compared to the Weakest-link model, A∗A∗- and V∗V∗-model.•The new model show good prediction capability for uniaxial tests specifically.•The loaded volume and the gradient effects are reflected with the new model.

In this paper, new probabilistic HCF evaluation models based on the surface and the sub-surface region is presented. The sub-surface model stem from the hypothesis that the highest fatigue risk depends on the loading in a region slightly below the free surface area. There, the materials microstructural barriers must be overcome by fatigue crack propagation. The size of the region with high load below the surface affects the fatigue risk. The predictive capability of the new proposed HCF model is studied along with comparisons to other state of the art HCF probabilistic models. Fatigue data originating from different geometries and materials conducted at different loading modes, rotating bending and uniaxial loading is used for the analysis. The investigation performed in this work indicate that the new proposed models increase the prediction accuracy and hence predictive capability.