About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Melt Processing, Casting and Recycling
|
| Presentation Title |
Automated Metal Cleanliness Analyzer (AMCA): A Multi-Class Framework for Inclusion Classification in Aluminum Melt Analysis |
| Author(s) |
Hannes Zedel, Mårten Görnerup, Robert Karl Fritzsch, Ragnhild Elizabeth Aune |
| On-Site Speaker (Planned) |
Hannes Zedel |
| Abstract Scope |
Assessing inclusion content in aluminum melts is essential for ensuring product quality in aluminum production. Traditional methods such as PoDFA (Porous Disk Filtration Apparatus) provide detailed cleanliness assessments but rely on manual analysis, limiting throughput and introducing subjectivity. Deterministic image analysis provides a fast and transparent alternative.
We present an expansion of the Automated Metal Cleanliness Analyzer (AMCA), building on previous work that demonstrated strong correlation with PoDFA-derived Total Inclusion Counts (TIC). The updated classification logic enables automatic detection of five TIC-relevant inclusion groups—Carbides, Grain Refiners, Refractory phases, Spinel, and Mixed Oxides—alongside three non-TIC categories: Filter Grain, Aluminum Bulk, and Intermetallic or artifactual structures.
AMCA’s classification logic utilizes statistical analyses of manually annotated inclusions for deriving deterministic rulesets. Our results demonstrate strong agreement with PoDFA-derived cleanliness metrics. This advancement positions AMCA as a scalable, transparent, and cost-effective alternative for process monitoring and quality control in aluminum production. |
| Proceedings Inclusion? |
Planned: Light Metals |
| Keywords |
Aluminum, Characterization, Recycling and Secondary Recovery |