About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Aluminum Reduction Technology
|
| Presentation Title |
THERMAL EVENTS CHARACTERISATION BY STUDYING SHELL TEMPERATURE IN ALUMINUM ELECTROLYSIS CELL |
| Author(s) |
Bazoumana Sanogo, Lukas Dion, Sébastien Gaboury, Sébastien Guérard, Jean-François Bilodeau, László Kiss |
| On-Site Speaker (Planned) |
Bazoumana Sanogo |
| Abstract Scope |
Maintaining thermal balance in aluminum electrolysis cells is essential for operational stability, energy efficiency, and process longevity. This balance is affected by thermal events from normal operations and anomalies such as anode effects. A key indicator is the spatial and temporal evolution of the ledge which is highly sensitive to thermal fluctuations.
While previous studies focused on shell temperature and ledge tracking, few have addressed characterization and detection of thermal events. This study aims to characterize thermocouple signals associated with thermal events, particularly anode changes.
Using a combination of signal processing and machine learning, specifically k-means clustering, distinct thermal patterns were identified. Data were collected from a fully instrumented industrial cell and processed through feature extraction and selection.
The findings were validated by an analog model and a full-scale industrial cell. This method demonstrated reliable pattern recognition and opens new avenues for process control, predictive maintenance, and operational efficiency. |
| Proceedings Inclusion? |
Planned: Light Metals |
| Keywords |
Aluminum, Machine Learning, Other |