This work pursues computational micromechanics approaches that define and compute mesoscopic Fatigue Indicator Parameters (FIPs) which serve as surrogate measures of driving forces for fatigue crack formation and microstructurally small crack growth. Attention is focused on constructing the extreme value distributions of FIPs as a function of microstructure to facilitate relative rank-ordering of fatigue resistance of microstructures as a function of thermomechanical process-history for a given alloy composition. Data science correlations are considered as a means to reduce uncertainty associated with model forms and parameters and to accelerate assessment of hot spot FIP distributions used to rank order microstructures. Further extensions in fusing information obtained from such computational strategies with in situ measurements of microstructurally small cracks and AI-enhanced crack detection methods for high cycle fatigue (HCF) are explored.
EXTENDED ABSTRACT