Complas 2025

COMPLAS 2025

Keynote

A data driven multiscale framework for understanding hardening and fatigue of alloys

Cui, Yinan (Tsinghua University)
Liu, Xin (Tsinghua University)
Liang, Zhun (Tsinghua University)
Tan, Fusheng (Tsinghua University)
Martinez Saez, Enrique (Clemson University)
Po, Giacomo (University of Miami)

In session: 706A - Machine Learning Empowered Computational Plasticity, Fracture and Interdisciplinarity I

Please login to view abstract download link

Although numerous efforts are spent, how does the nanoscale compositional heterogeneity and microscopic mechanism of alloys control its macroscopic performance remains far from well understood. In the current work, a new data-driven multiscale framework is developed, which coupled the high-throughput molecular dynamics (MD) simulations and larger scale discrete dislocation dynamics (DDD), and crystal plasticity theory (CP), to explore the compositional heterogeneity controlled hardening behavior of high-entropy alloys (HEAs), as well as the micromechanism controlled fatigue behavior of superalloys. The atomic arrangement influenced fundamental dislocation behaviors under spatially non-uniform stacking fault energy across the slip planes are studied. The fatigue life prediction model is proposed. The framework paves the way for designing alloys with tailored macroscopic properties by manipulating their local structural attributes and microstructure features.