Analysis of Influence Factors on the Melting Point of the Freeze Slag Inside Flash Smelting Furnace Brickless Reaction Shaft

"Brickless reaction shaft is the latest progress for flash smelting furnace. However, there are many factors influence the melting point of the freeze slag inside brickless reaction shaft, and their variational regularities are not obvious, thus, it is difficult for conventional methods to analyze accurately the complex data. In this research, the experiment data of slag melting point were processed based on optimal discrimination plane. For this method, two orthogonal vectors were firstly built up based on the Fisher's criterion, and then the experiment data were projected onto the two vectors, thus, two-dimensional feature vectors were extracted as criteria to determine the factors effect degree on the slag melting point. Results show that the CaO content is the greatest factor to the melting point of the freeze slag inside brickless reaction shaft, the content of Fe3O4, SiO2, Cu2 O and MgO is the second, and the FeO content is the last.IntroductionBrickless reaction shaft is the latest progress of flash smelting furnace reaction shaft. To research the influence factors on the slag melting point inside the brickless reaction shaft is significant to form freeze slag with compact structure, good heat insulation, certain intensity, appropriate thickness and reasonable distribution, to extend furnace life, to reduce energy consumption, and to promote safe production.However, due to the complex components, there are many factors influence the melting points of the freeze slag inside brickless reaction shaft, and their variational regularities are not obvious. Therefore, it is difficult for conventional methods, such as variance analysis, data fitting, etc., to analyze accurately the complex experiment data.Pattern recognition can easily classify information, even if there are many influencing factors. For pattern recognition, there are some sub-disciplines such as discriminant analysis, feature extraction, error estimation, cluster analysis (together sometimes called statistical pattern recognition), grammatical inference, parsing and matching (sometimes called syntactical and structural pattern recognition) [1, 2]."