/VQ-SM

Primary LanguagePython

VQ-SM: Surrogate model based on VQ-VAE for damage forecast of CFRP subjected to LVI

  • A framework of using advanced representation learning technique VQ-VAE to enhance the performance of the surrogate model.
  • The forecast of full-field delamination and the matrix damages of CFRP are investigated subjected to LVIs of different impactor radius $R$, impact energy $E$ and the impact angle $\theta $.

Environment

  • python39
  • pytorch (torch 1.12.1+cu116 used)

Dataset and Pre-trained models

The dataset and the pre-trained model are available from Google Drive

Evaluation using VQ-SM for damages

Training a VQ-SM