Research on equipment portrait technology of aluminum electrolyzer with improved spectral clustering

Electrolyzer is an important equipment to produce aluminum in the aluminum electrolysis industry. The change of production index in the process of electrolysis will affect the current efficiency to different degrees. By constructing the portrait of the electrolytic cell, it can provide research ideas for analyzing the state of the electrolytic cell and mining the relationship between production indicators. This paper takes the electrolytic cell as the research object, uses the production data in the electrolysis process, combines the portrait technology, clustering and other methods, extracts the label information, and constructs the portrait of the electrolytic cell to facilitate the analysis and decision-making of the staff. Aiming at the difficulty of selecting the optimal scale parameter in spectral clustering algorithm, an improved adaptive spectral clustering algorithm based on local coefficient of variation was proposed to improve the clustering performance. Experiments are carried out through actual production data to generate portraits of electrolyzer equipment, which provides reference for practical applications.