A novel knowledge-driven intuitionistic fuzzy multi-criteria decision-making method for evaluating the operating condition of the aluminum electrolysis cell

Background: Accurate evaluation of aluminum electrolysis cell operating conditions is essential for maintaining safe and stable production processes. However, the limited number of manually annotated operating condition labels, derived from various process indicators, hinders the effectiveness of prediction models across diverse datasets of aluminum electrolysis cells.
Method: This study proposes a novel knowledge-driven intuitionistic fuzzy multicriteria decision-making (MCDM) method, which utilizes expert-defined aluminum electrolysis cell operating standards to evaluate operating conditions without dependence on labels. By leveraging expert-defined operating standards for aluminum electrolysis cells, unified decision matrices that integrate process indicators with these predefined standards are constructed. To determine optimal indicator weights, Intuitionistic Multiplicative Preference Relations (IMPRs) with the Criteria Importance Through Inter-criteria Correlation (CRITIC) method are combined together, effectively balancing subjective expert knowledge and objective data-driven insights. Finally, the fuzzy Elimination and Choice Translating Reality (ELECTRE) II method is utilized to create comprehensive ranking relationships between operating standards and process data, thereby improving the accuracy and reliability of operating condition evaluations.
Significant findings: Case studies validate the effectiveness of the proposed method in evaluating operating conditions.
© 2025 Taiwan Institute of Chemical Engineers.