Revealing strain partitioning and GNDs-mediated hardening in heterostructure low-alloy steel: An in-situ EBSD validation of a bending simulation methodology
Jiazheng Zhao, Shengen Zhang, Jian Wang, Jun Zhang, Jun Li, Fenghua Luo
Abstract
• In-situ EBSD validation of bending-based strain simulation for heterostructures steel. • Accurate strain prediction (MAE < 10 %) in recovered/deformed bimodal structures at low strains (<10 %) • Defined validity domain: reliable for GND-dominated heterostructures excluding necking stage. To investigate and optimize the strength-ductility of heterostructure low-alloy steels in a cost-effective manner, in-situ EBSD is employed to examine the strain partitioning behavior and microstructural evolution of typical recovery/deformation dual-mode and recovery/deformation/recrystallization multimodal heterostructure steels during uniaxial tension. It is clarified that the plastic strain preferentially initiates within the softer recovered grains and propagates from the grain boundaries inward. The harder deformed grains subsequently accommodate strain for compatibility, whereas fine recrystallized grains exhibit limited strain-bearing capacity and contribute minimally to later-stage plasticity. The significant strain gradient at the recovered/deformed grain interfaces is crucial for inducing a massive accumulation of GNDs, which generates a pronounced hetero-deformation induced hardening effect. Compared to the multimodal heterostructure, the dual-mode heterostructure accumulates GNDs more sustainably and efficiently, leading to superior work hardening and ultimately achieving a yield strength of 527 MPa and an elongation of 14.2 %. The first systematic validation of the proposed “bending simulation” method, quantifying its effectiveness, applicable boundaries, and error margins through in-situ experiments. The bending test is confirmed to be feasible for simulating grain-level strain partitioning during the yielding and early work-hardening stages (total error less than 10 %), offering an efficient and economical alternative for performance prediction and microstructural design.