Investigation Of Magnesia-carbon And Alumina-zirconia-carbon Slide Plate Refractories

With the development of technique and process of fast and efficient continuous casting and secondary refining, slide gate system becomes more and more important, slide plate as the most important part of the system is now extensively used to tap the molten steel from basic oxygen furnace and control the flow rate from ladle and tundish. In the course of casting, it should choose appropriate slide plate refractories when casting different type of steel, for example, magnesia-carbon (MgO-C) or magnesia-spinel slide plate is applied to cast calcium-treated steel, and alumina-zirconia-carbon (AZC) slide plate is used for casting ordinary steel.Magnesia-carbon and alumina-zirconia-carbon slide plate refractories were investigated in chapter two and three of this thesis, respectively. The detailed research work was carried out as follows:Firstly, the influence of the content of metallic aluminum powder on the physical properties of MgO-C materials treated at different temperature in air was investigated, and then the phase composition and microstructure were analyzed, and the regularity of the effect of aluminum powder on the structure and properties was discussed to determine its appropriate content.Secondly, the influence of graphite and spinel fines on the properties of MgO-C materials was investigated. The results show the strength declines with increased content of graphite, and aptly added some spinel fines can suppress the aluminothermic and carbothermic reducing reactions of magnesia.Thirdly, the effect of single silicon powder on the structure and properties of coked AZC materials was investigated. The results show silicon powder with middle particle size can increase the strength remarkably, but that with large particle size can improve the thermal shock resistance.Fourthly, orthogonal experimentation was adopted to research the effect of multiplex silicon powder on the properties of coked AZC materials, and the excellent experimental recipe was choosed after the analysis of range and variance.Finally, Data mining method (Parital least square, artificial neural network, and support vector machine) was used to build the appropriate mathematic model of crushing strength before and after thermal shock, and then predict the properties of the choosed recipe. The recipe satisfied the industry demand can be further determined after the validation experiment.