New P-DivGNN framework uses graph neural networks for advanced stress field reconstruction

Researchers have developed P-DivGNN, a novel physics-informed machine learning framework designed to reconstruct local stress fields in micro-scale simulations. This approach utilizes a graph neural network to represent micro-structures and incorporates physical constraints to ensure stress equilibrium and enforce periodic boundary conditions. The P-DivGNN framework demonstrates significant computational speed-ups compared to traditional finite element methods, particularly for non-linear hyperelastic responses, making it a promising tool for large-scale applications in areas like fracture analysis and fatigue criteria definition. AI

IMPACT Accelerates micro-scale simulations for fracture analysis and fatigue criteria, offering significant speed-ups over traditional methods.

RANK_REASON The cluster contains a research paper detailing a new physics-informed machine learning framework. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• finite element method
• graph neural network
• Hugging Face
• machine learning
• Manuel Ricardo Guevara Garban
• P-DivGNN

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →