Simulation-based extreme value marked correlations in fatigue of advanced engineering alloys

Abstract

The objective of this research is to estimate the microstructure dependence of the extreme value probabilities of fatigue crack formation at the scale of the grains in polycrystalline and polyphase microstructures to facilitate preliminary parametric design exploration and property assessment. A simulation-based methodology is introduced for computing correlation functions of microstructure attributes marked by the extreme value microstructure-scale fatigue indicator parameters. Multiple statistically representative microstructure volumes simulated using a crystal plasticity constitutive model coupled with the FEM are used to characterize these extreme value marked correlation functions. By comparing these marked correlation functions to traditional correlation functions sampled from the bulk material, the interacting microstructure attributes important to the extreme value response are identified. This methodology is applied to identify dominant crystallographic relationships relative to fatigue crack formation in the polycrystalline P/M Ni-base superalloy IN100 and duplex Ti-6Al-4V. It is demonstrated that certain high contrast orientation relationships in both IN100 and Ti-6Al-4V (where grains oriented for hard slip are near grains oriented for easier slip) can significantly increase the driving forces for fatigue crack formation.