Anisotropic plasticity identification: Integrated FE and sensitivity-based virtual fields method

Abstract

An integrated virtual field method (VFM) framework has been developed for the first time by integrating finite element-based VFM (FE-VFM) and sensitivity-based VFM (S-VFM) approaches. This new methodology enhances computational efficiency while maintaining high accuracy in the inverse identification of constitutive model parameters. Full-field displacement measurements are mapped onto finite element meshes to improve data quality and reduce the number of measurement points. Sensitivity-based automatic virtual field generation is incorporated to robustly handle complex anisotropic plasticity models. The proposed approach has been validated using both noise-free virtual experimental data and actual experimental data from SUS316 stainless steel. Results demonstrate that Swift hardening law and quadratic Hill-48 yield function parameters can be accurately identified from a single specimen, achieving significantly reduced computational time compared to conventional VFM. Furthermore, for the first time, the non-quadratic Yld2000–2d anisotropic yield function parameters have been successfully identified using a single specimen with an additional cost function incorporating balanced biaxial tension data. This integrated VFM represents a novel and efficient inverse identification framework, enabling robust determination of complex anisotropic constitutive models with minimal testing.