XDEM: Physics-Informed AI Framework for Crack Propagation
bravo_abad · x · 2026-08-24
Yizheng Wang and coauthors introduce XDEM (Extended Deep Energy Method), a physics-informed neural framework for fracture mechanics.
- Challenge: Fracture is difficult for SciML because solutions contain discontinuities. Standard approaches require prior knowledge of crack geometry or dense collocation points around crack tips.
- Method: XDEM incorporates physics more directly. For discrete fracture, the neural representation explicitly includes a crack function for displacement discontinuities and functions encoding asymptotic behavior near crack tips. For continuous fracture, it couples neural representations.
- Significance: It enables solving fracture problems without needing precise prior knowledge of crack geometry.
More from Research
- Open-source Exo framework enables agent self-modification and time travel rollback — omarsar0 · 2026-08-24
- InfinityEdit enables infinite video editing with a lightweight adapter — _akhaliq · 2026-08-24
- 6,000 fruit picks: ChatGPT's "random" choice changes sharply by language — Raffallos · 2026-08-24
- Bonsai: Tree-Based Method Preserves High-Dimensional Data Structure Better Than UMAP — bravo_abad · 2026-08-24
- Are We Smart Enough to Evaluate Model Intelligence? — CSProfKGD · 2026-08-24
- Interactive site explains SynthID text watermarking: algorithm and six costs — neil_chilson · 2026-08-24