 |
Approach for learning constitutive laws in the context of experimental mechanics
1 : Laboratoire de Mécanique Paris-Saclay
(LMPS)
-
Site web
CentraleSupélec, Université Paris-Saclay, Centre National de la Recherche Scientifique, Ecole Normale Supérieure Paris-Saclay
4 avenue des sciences / 8-10 rue Joliot Curie, 91190 Gif-sur-Yvette -
France
To design a structure, it is essential to have a reliable constitutive law. When using neural network-based modeling, this reliability depends on the data selected for training, raising the question of which experiments should be performed. Which loadings should be chosen to provide the most informative data? One approach is to define, a priori, a quality measure based solely on the observations and correlated with the model's performance. This measure can then serve as an indicator to guide the collection of experimental data.
