.png){kind=icon}
.png){kind=icon}
Return
Predicting plastic strain localization in porous solids using graph neural networks
.png){kind=icon}
OA
AI
DOI：10.1016/j.mechmat.2026.105733.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Graph neural networks are used to predict strain localization in 2D porous solids. • Graph nodes represent the voids and graph edges represent potential shear bands. • Data-driven approach provides high fidelity, but requires extensive training data. • Hybrid approach augments the neural network with physical prior from limit analysis. • Hybrid approach retains macroscopic accuracy, needs less data and generalizes better.
Keywords:
Surrogate modeling
Graph neural network
Plasticity
Heterogeneous material
Porous solid
Strain localization
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.1
Papers:
4.0K
Citations:
1.2W