| 초록 |
In this presentation, the multiscale mechanical modeling of polymer nanocomposites and data-driven computational mechanics framework will be introduced. The precise characterization of interphase is critical in the prediction of the effective elastic modulus, strength and toughness of the polymer nanocomposites. To achieve this purpose, the sequential multiscale bridging method (based on the molecular dynamics and the finite element homogenization) has been developed. However, the nano-micro-macro concurrent multiscale analysis was still challenging issue because the molecular dynamics needs large computational time. In this study, to overcome the limitation, the molecular dynamics data-driven computational mechanics framework based on the neural network is proposed. Furthermore, the proposed approach will be applied to the prediction of the fracture toughness of polymer nanocomposites. |
