|
| Study on Data Analysis and Visualization of Lifespan of Aluminum Electrolysis Cells |
|
Received:September 11, 2024
Revised:September 11, 2024
Accepted:September 12, 2024
Published Online:October 29, 2024
|
| View Full Text View/Add Comment Download reader |
| DOI：doi：10.3969/j.issn.1007-7545.2024.11.012 |
| KeyWord:aluminum electrolysis; cell lifespan; exploratory data analysis; distributed sensing |
| Author | Institution |
| LI Jie |
中南大学 冶金与环境学院 |
| SHI Huihong |
中南大学 冶金与环境学院 |
| PENG Chen |
微软中国有限公司无锡分公司 |
| CHEN Can |
中南大学 冶金与环境学院 |
| CHEN Kaibin |
中铝郑州有色金属研究院有限公司 |
| LUO Yingtao |
中铝郑州有色金属研究院有限公司 |
| ZHANG Yanan |
中铝郑州有色金属研究院有限公司 |
| WANG Huaijiang |
中铝郑州有色金属研究院有限公司 |
| ZHANG Hongliang |
中南大学 冶金与环境学院 |
|
| Hits: 1522 |
| Download times: 847 |
| Abstract: |
| With the deepening development of industrial informatization, the volume of data generated during aluminum electrolysis production has increased significantly, rendering traditional data analysis methods insufficient for addressing complex industrial environments. In this paper, a big data analysis framework for aluminum electrolysis based on an edge-cloud real-time sensing system was proposed, which integrates online monitoring, manual offline detection and recorded data. Firstly, the classical statistical analysis methods were employed to conduct an initial study on the lifespan and process parameters of aluminum electrolysis cells, revealing an average lifespan of 2 064 days, with 58% of the cells operating for over 2 000 days. Then, the Exploratory Data Analysis (EDA) method was applied to perform data preprocessing, including handling missing values, noise reduction, and feature scaling. The distribution characteristics and correlations of process parameters are visualized. Finally, the key variables were extracted, which provides foundational data for the establishment of the lifespan prediction model for aluminum electrolysis cells. |
| Close |
|
|
|