Data-driven material modeling using Constitutive Relation Error
Author Information
Author(s): Ladevèze Pierre, Chamoin Ludovic
Primary Institution: CentraleSupélec, ENS Paris-Saclay, CNRS, LMPS-Laboratoire de Mécanique Paris-Saclay, Université Paris-Saclay
Hypothesis
What is the mathematical form of the most general data-driven constitutive model for stable materials?
Conclusion
The study presents a data-driven approach to learn constitutive models for stable materials, focusing on elasto-(visco-)plastic materials.
Supporting Evidence
- The proposed method leads to quasi-explicit formulations of the optimal constitutive model.
- A modified Constitutive Relation Error is introduced to account for measurement noise.
- The study focuses on stable elasto-(visco-)plastic materials under small displacement assumptions.
- The CRE framework can be used to calculate data-driven models defined by thermodynamic potentials.
Takeaway
This study helps scientists understand how to create models for materials using data, which can make predicting how materials behave easier.
Methodology
The study proposes a general data-driven approach to learn constitutive models from experimental data using the Constitutive Relation Error (CRE) as a minimization tool.
Limitations
The study does not cover numerical treatment involving optimization and regression problems.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website