About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Aluminum Electrode Technology Symposium
|
| Presentation Title |
Rhodax® Granulometry Size Distribution Analysis for Green Anode Density Optimization via Machine Learning Cluster-Based Method |
| Author(s) |
Louna Alsouki, Christophe Bouche, Vincent PHILIPPAUX, Fabienne Virieux |
| On-Site Speaker (Planned) |
Louna Alsouki |
| Abstract Scope |
This study undertakes the impact of anode aggregate size fractions produced by the Rhodax® grinding process on green anode quality using data science techniques. Industrial data were collected, containing aggregate size distributions and the corresponding green anode density values. Since aggregate fractions constitute compositional data, centered log-ratio (CLR) transformations were applied as a preprocessing step to enable use of conventional statistical and machine learning methods while preserving the relative nature of the data.
An unsupervised clustering approach was implemented to classify granulometry size distributions and identify distinct grinding modes hidden within the dataset. All cumulative particle size distributions were analyzed and correlated to their associated green anode densities to determine the granulometric characteristics linked to best anode quality.
The results provide data driven insights into the relationship between granulometry size distribution and green anode density and help in identifying the most favorable grinding procedure for optimizing anode quality. |
| Proceedings Inclusion? |
Planned: Light Metals |
| Keywords |
Process Technology, |