Operational Status Evaluation And Non-Optimal Reasons Traceability For The Smelting Process Of Fused Magnesium Furnace

As China's industrial development is accelerating,the problem of energy depletion is becoming more and more serious.Improving energy efficiency is in line with the long-term interests of human development.In the industry,energy-saving and emission reduction of the high-energy consumption processes such as fused magnesia furnace has become the development trend of modern industry.In order to ensure the economic benefits of refining magnesia,it is particularly important to optimize the operating state for the smelting process of fused magnesium furnace(FMF).It has important theoretical value for improving energy utilization and improving enterprise competitiveness to realize the on-line evaluation of the operating state of the FMF,and to trace the non-optimal reasons when the operating state is non-optimal for guiding the subsequent production process.The evaluation of the operational status of the production process involves many methods,such as case-based reasoning,method based on gray correlation analysis,principal component analysis,neural network,and Fisher discriminant analysis methods.In this paper,for some features of FMF such as less measurable variables,the coexistence of qualitative information and quantitative information,multi-production conditions,multiple processes alternately and continuously,process variable coupling,etc,the method based on CBR for the evaluation of the operating state of the multi-condition and complex process is proposed,and it is realized the online evaluation of the operating state of the smelting process of the FMF and the traceability of the non-optimal state.It is firstly analyzed the theoretical model of single-ton energy consumption in the FMF,and selected the key process variables that can be used for operational state evaluation in this paper.For the features of less measurable variables and the coexistence of qualitative information and quantitative information,through in-depth analysis of the process flow of each process and combining with the actual process,the auxiliary evaluation variables that can be used for operational status evaluation are selected.For the feature of multi-production conditions,the method of dividing the working conditions based on qualitative information classification is proposed.For the feature of multiple processes,a process-level classification method based on similarity is proposed.The model based on CBR for the evaluation of the operating state of the multi-condition and multi-process complex process is established,which can achieve the online evaluation of the operating state of the FMF.For the non-optimal operating state,the method based on variable contribution rate has found the main variables that lead to the non-optimal operation state,and achieved positioning of non-optimal causes.Finally,the effectiveness of the above methods is verified by simulation tests.