COMPLAS 2025

Multiscale modeling strategy for predicting fatigue life of steel based on crack growth simulation

  • Shibanuma, Kazuki (The University of Tokyo)

Please login to view abstract download link

This study presents a multiscale modelling strategy for accurately predicting the high-cycle fatigue strengths of steels. The required input data are only microstructural information, tensile properties, and loading conditions, without any adjustable material constants. In the proposed strategy, total fatigue life is estimated from crack growth life alone. The entire model comprises three sub-models, for: (i) a macroscopic finite element analysis, (ii) microstructure, and (iii) crack growth. The model was strictly validated against the results of experiments performed on three different steels under various loading conditions. Although the experimental fatigue life results exhibited wide variation, the predicted and experimental data were accurately matched over the entire range. The notch sensitivity of the fatigue limits depending on the material strength were successfully reproduced by the proposed model. The transition of the crack growth rate was also accurately predicted compared with the experimental results. The results demonstrate that the fatigue life of steels under high-cycle fatigue can be accurately predicted from crack growth life alone. Furthermore, the proposed strategy is capable of effectively explaining the dependence of fatigue strength on microstructure and loading conditions based on the fracture mechanics.