3.6: Data-driven methods for inelastic solids and structures

Organizers:

  • Yupeng Zhang, Caltech
  • Burigede Liu, University of Cambridge
  • Miguel A. Bessa, Brown University

Description:

The inelastic response of solids and structures can be induced by a wide range of factors, such as finite deformation, elevated temperature, high strain rate loading, and other extreme environmental conditions. These inelastic behaviors are essentially history-dependent and often computationally intensive to model due to the numerical efforts required for convergence and the potential for instability. With advancements in computational capabilities, data-driven methods, such as deep learning and statistical approaches, have been developed to address both fundamental and applied problems in inelastic solids and structures.


Topics of interest:

This mini-symposium welcomes all relevant submissions, including but not limited to:

  • Model reduction, such as homogenization of constitutive relations, and further for multiscale modeling.
  • Inelastic behaviors of solids, such as (crystal) plasticity, damage, brittle and ductile fracture.
  • Inelastic behaviors of structures, such as architected metamaterials.
  • Inverse problems, such as characterization of inelastic material properties.
  • Frameworks and algorithms of data-driven methods, such as algorithms for operator learning.
  • Transfer learning, such as fine tuning of models for out-of-distribution data.