Calibration of thermo-viscoplastic constitutive model under biaxial loadings: A Feasibility Study
Résumé
In this study, an inverse identification strategy based on the finite element model updating (FEMU) method is proposed to calibrate the parameters of a temperature and strain rate dependent constitutive model. The mechanical responses of a dedicated cruciform specimen with heterogeneous temperature field under biaxial loading are employed to supply information to the inverse scheme. A combination of Particle Swarm Optimization (PSO) and SIMPLEX optimization algorithm is employed to find the optimal values of the material parameters. In order to validate the proposed identification strategy, a virtual experiment is designed and performed with a reference material constitutive model. The proposed strategy is proved to be feasible as all seven parameters of the constitutive model are accurately identified. In addition, the influences of measurement noise of force, temperature, and strain data are analyzed by means of a sensitivity study. The experimental data after the localized necking should be avoided for parameter identification. The proposed inverse identification strategy shows good robustness to strain noise.
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Licence : CC BY - Paternité
Licence : CC BY - Paternité