Numerically predicted high cycle fatigue properties through representative volume elements of the microstructure

Abstract

The investigation of high cycle fatigue (HCF) properties of materials is elaborate in experimental design. To reduce the costs, a multiscale numerical approach to predict the HCF-strength is proposed in this study. The intention is to correlate microstructural features and their related microdeformation mechanisms with macroscopic fatigue properties. The approach involves a series of numerical and mechanism-based analytical models. The microstructural features, including phase fraction, grain size and grain shape, are statistically characterized and represented in the generated representative volume element (RVE) model of the microstructure of a two phase steel. The underlying microdeformation mechanism is captured by an extended crystal plasticity (CP) model incorporating kinematic hardening on slip systems for cyclic loadings. The CP parameter set was calibrated on strain controlled low cycle fatigue tests. The numerical simulations of the cyclically loaded RVEs resulted in local plasticity fields. The highest plastically deformed grain of each RVE was identified and its grain size averaged plastic strain value extracted for further analysis. These scattered plasticity values fit extreme value distribution density functions. The parameters, which describe the shape of the functions, were used to calculate the HCF-strength. The obtained values showed a good agreement to experimental results.