Data-Augmented Modeling Of Constitutive Laws for Engineering Systems

The modeling of complex engineering systems is highly challenging. Physics-based models require a cautious application of constitutive assumptions, whereas data-based models require vast amounts of data. DAMOCLES (Data-Augmented Modeling Of Constitutive Laws for Engineering Systems projects) targets a breakthrough in the constitutive modeling of such systems in different physical domains by developing a unified multi-tool framework that combines the favorable characteristics of physics-based and data-based approaches. As one part of the DAMOCLES project, this research aims to learn constitutive laws in a multi-physics setting using an ANN-based Hamiltonian model learning framework.

Existing highly flexible data-driven model learning methods can capture a wide range of system behaviors, but lack good generalization properties, and the resulting models are difficult to interpret due to their black-box character. Hence, this project aims to use an ANN-based Hamiltonian model learning framework to obtain interpretable, robust, and generalizable models of the underlying system dynamics.