About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
High Performance Steels
|
| Presentation Title |
Automated ASTM-E45 Inclusion Classification Using Deterministic Image Analysis |
| Author(s) |
Hannes Zedel, Mårten Görnerup, Ragnhild Elizabeth Aune |
| On-Site Speaker (Planned) |
Hannes Zedel |
| Abstract Scope |
Accurate assessment of non-metallic inclusions is essential for steel quality assurance. ASTM E45 provides the established framework, yet its manual application remains the industrial norm. Manual ratings are labor-intensive, subjective, and difficult to scale, but remain entrenched in certification and contractual requirements.
We demonstrate a proof-of-concept hybrid solution that combines deterministic digital image analysis with streamlined human review. The pipeline automates illumination correction, segmentation, hot-axis estimation, clustering, and severity mapping in line with ASTM E45 rules. Results are presented as graphical overlays and software-based tables that allow operators to rapidly review and correct features before final reporting.
This hybrid approach preserves contractual compliance while reducing bias and workload. It demonstrates the feasibility of reproducible, high-throughput cleanliness assessments that generate statistically robust datasets, creating value both for quality assurance and for process and product development. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Iron and Steel, Characterization, Computational Materials Science & Engineering |