| Summary: | A microstructure-based fatigue life prediction methodology was proposed and verified for 44MnSiVS6 hypo-eutectoid steel. As a first step, a statistically equivalent synthetic microstructure generation algorithm was developed by applying the cellular automata under the multi-level microstructure generation approach, and verified by comparing to the real microstructure. Crystal plasticity-based elasto-viscoplastic constitutive models were developed for the primary ferrite and pearlite, and their constitutive parameters were calibrated. Synthetic microstructure-based crystal plasticity finite element method (CP-FEM) simulations were performed for stress-controlled axial fatigue tests. The Fatemi-Socie (FS) parameter was chosen to calculate a fatigue indicator parameter increment (ΔFIP) per cycle, and its cumulative distribution probability (CDP) curve was obtained from CP-FEM simulations. Three different criteria extracting a representative ΔFIP from the CDP curve were proposed after thoroughly examining the nature of the CDP curves, which showed the dual transition in the upper CDP range. Extracted representative ΔFIP values coupled with the power law-based model gave reasonable fatigue life prediction. In particular, the best prediction was made when the representative ΔFIP values were taken beyond the second transition of the CDP curve, in which the ΔFIP hot spots were concentrated in the primary ferrite.
|
|---|