Graph-based representation of history-dependent material response in the Data-Driven Computational Mechanics framework - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Computer Methods in Applied Mechanics and Engineering Année : 2023

Graph-based representation of history-dependent material response in the Data-Driven Computational Mechanics framework

Résumé

This work develops an extension of the Data-Driven Computational Mechanics approach introduced by Kirchdoerfer and Ortiz (2016) to history-dependent behaviour. The material database is replaced by a material directed graph, where vertices represent strain-stress pairs and arcs encode thermomechanically admissible transitions between them. A graph search algorithm is used to select local material databases representative of the incremental material behaviour. This study thus gives the mathematical tools for a structured data-driven representation of the material behaviour, designed to minimise the increase of the problem dimensionality. The performances of the proposed approach applied to elastoplasticity are studied on a single-element structure and a larger truss structure. A two-step solver is proposed to address numerical accuracy issues in the latter case.
Fichier principal
Vignette du fichier
Dandin_Leygue_Stainier_preprint_v2.pdf (2.93 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04259707 , version 1 (26-10-2023)
hal-04259707 , version 2 (01-12-2023)

Identifiants

Citer

Héloïse Dandin, Adrien Leygue, Laurent Stainier. Graph-based representation of history-dependent material response in the Data-Driven Computational Mechanics framework. Computer Methods in Applied Mechanics and Engineering, 2023, 419, pp.116694. ⟨10.1016/j.cma.2023.116694⟩. ⟨hal-04259707v2⟩
86 Consultations
74 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More