The phase-field fracture (PFF) model is attracting attention as a powerful tool for predicting complex fracture modes, but its accuracy relies on the accurate calibration of material parameters. On the other hand, the digital image correlation (DIC) method, which enables full-field measurement on the target surface through image analysis techniques, is utilized to capture deformation and fracture phenomena in various materials. This study proposes a robust framework to estimate parameters of the PFF model for ductile fracture based on a Bayesian data assimilation method that utilizes the strain field on the surface captured using the DIC method. To validate the proposed method, we conducted a uniaxial tensile test on a notched A5052-0 specimen and inversely calibrated the critical energy release rate. The proposed method successfully identified an optimal parameter by minimizing a cost function based on the difference between the experimental and simulated strain distributions. The validity of the estimated parameter was confirmed by a PFF simulation, which demonstrated a qualitative match in fracture behavior and a good quantitative agreement in the force-displacement response compared to the experiment. This work demonstrates the effectiveness of combining DIC measurements and data assimilation for calibrating PFF models.
SASAKI et al. (Wed,) studied this question.