Operation Performance Assessment and Non-optimal Reasons Traceability for Melting Process of Fused Magnesium Furnace

A method of operation performance assessment based on the combination of fuzzy classification and Gaussian mixture model is proposed to assess the operation performance for the smelting process of electrofused magnesium furnace. And a method of non-optimal cause tracing based on the combination of contribution diagram and case-based reasoning is used to trace the non-optimal causes. For the qualitative data, the fuzzy classification method is used to transform the qualitative division into the quantitative division. Then, the posterior probability of the Gaussian mixture model of each running state of the online data is calculated to get the evaluation results. When the evaluation result is not optimal, the contribution rate of each variable is compared to obtain the non-optimal variable, and the non-optimal reason is found by case search. Finally, the effectiveness of the proposed method is verified by the simulation platform of electrofused magnesium furnace.