画像計測データとフェーズフィールド破壊モデルによるき裂進展現象のデータ同化

抄録

Although various methodologies for numerical simulation of crack propagation have been developed to accurately predict fracture behavior in mechanical structures, the phase-field fracture (PFF) model is a powerful simulation methodology. The PFF model has the advantage in simulating complex crack path by introducing an auxiliary variable called the phase-field variable, which represents the fracture level at each coordinate. A crack propagation is simulated by minimizing the potential energy defined as a function of the phase-field variable. On the other hand, data assimilation methodologies based on Bayesian inference have been attracting attention to integrate measurement data into numerical simulations. This study proposes to implement the data assimilation methodology to the PFF model and validate the capability of proposed method to estimate model parameters included in the PFF model through numerical experiments.