Finite element analysis has become an indispensable tool for optimizing mod-ern precision manufacturing processes, including aerospace component fabri-cation, semiconductor production, battery manufacturing, and automotive forming. Achieving reliable ...
Finite element analysis has become an indispensable tool for optimizing mod-ern precision manufacturing processes, including aerospace component fabri-cation, semiconductor production, battery manufacturing, and automotive forming. Achieving reliable and predictive simulations, however, requires ad-vanced material models capable of accurately describing complex defor-mation and failure mechanisms. Despite their effectiveness, such constitutive models typically demand extensive and labor-intensive experimental cam-paigns for parameter calibration, which limits their practical applicability.
To address this challenge, this dissertation proposes an integrated virtual fields method (VFM) framework for efficient and robust multi-scale material param-eter identification. At the macroscopic scale, a virtual experimental data–based identification strategy is first established using finite element analysis to sys-tematically calibrate parameters of the Swift hardening law and anisotropic yield functions, including Hill-48 and Yld2000-2D. The proposed integrated VFM enables reliable parameter extraction from a single heterogeneous spec-imen, significantly reducing experimental complexity. The approach is subse-quently validated using actual experimental measurements of SUS316, demonstrating improved accuracy and robustness compared with convention-al VFM formulations.
At the microscopic scale, the framework is further extended to crystal plastici-ty modeling through micro-digital image correlation (micro-DIC). Virtual ex-periments confirm the consistency and stability of the identification procedure, while experimental micro-DIC data obtained from IF and dual-phase steels are employed to determine physically meaningful crystal plasticity parameters for individual phases.
Overall, the proposed framework establishes a unified and scalable inverse identification methodology that can be consistently applied across length scales, enabling coherent and physically interpretable multi-scale material characterization. This integrated VFM framework constitutes the primary con-tribution and novelty of this dissertation.