Complas 2025

COMPLAS 2025

Homogenization of Mechanical Metamaterials

Ulloa, Jacinto (University of Michigan)
Ariza, Pilar (Universidad de Sevilla)
Andrade, José (California Institute of Technology)
Ortiz, Michael (California Institute of Technology)

In session: 704 - Data-Driven Methods and Machine Learning for Particulate Materials

Please login to view abstract download link

Multiscale methods play a significant role in the study of architected materials, with data-driven approaches gaining notable traction in recent years. Typically, these methods assume predefined kinematics and balance equations as universal principles, specifying the primary variables of interest a priori and incorporating material behavior through empirical data. We argue, in the context of mechanical metamaterials, that kinematics and balance equations require careful consideration. Importantly, the effective behavior of these materials exhibits generalized continuum mechanics, requiring non-standard primary variables for multiscale analysis. We will show that this effective behavior follows naturally from variational arguments.