About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Artificial Intelligence Applications in Integrated Computational Materials Engineering (AI-ICME)
|
| Presentation Title |
Autonomous Fractography Analysis Framework Driven by Large Language Models |
| Author(s) |
Quanliang Liu, Hyunseok Oh |
| On-Site Speaker (Planned) |
Quanliang Liu |
| Abstract Scope |
The increasing complexity of alloys, combined with diverse service environments, makes fracture analysis a significant bottleneck in both post-failure assessment and new alloy development. Traditional characterization of fracture surfaces relies heavily on the experience of human experts, which is subjective and limited in scalability. To address these challenges, this work introduces a framework that leverages multi-modal large language models (LLMs) to automate and enhance fracture surface analysis. We focus on fine-tuning multi-modal LLMs by developing a unique, domain-specific dataset of annotated fracture surface images from 3000 open-source papers. By training on this specialized data, the model develops visual and physics-informed reasoning ability to interpret complex fracture surfaces, identifying critical features and underlying mechanisms directly from fractography images. This work can serve as a crucial post-characterization module for automated alloy discovery pipelines, refining LLM-generated alloy design hypotheses by analyzing fractographs from mechanical tests and identifying weaknesses, enabling an iterative design-test-improve cycle. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
ICME, Machine Learning, Computational Materials Science & Engineering |