| Abstract:With China’s steel production increased year by year, the demand for iron ores are growing incessantly. Due to the increasingly tense of iron ore resources, the numerous varieties of imported iron ores and the uneven quality of domestic iron ores, the raw materials used in the sintering plant are becoming complicated. Instability of raw materials causes great difficulties to the granulation and sintering. Therefore, it has great significance to develop the Iron Ores Proportion Optimization System for Sintering, make the reasonable and efficient use of iron ores. And finally, stable sintering production, improve sinter quality, reduce production costs and even improve blast furnace technical and economic indicators.According to the characteristics of sintering process, we established the model of iron ores proportion optimization that use the blending cost as the objective function, the sinter chemical composition,-0.5mm particle content of mixture and supply conditions of iron ores for the constraints. Proposed the method of combining linear programming and genetic algorithm to solve the iron ores proportion optimization model, then obtain a group of iron ores blending schemes which meet the sinter chemical composition requirements, and has the better technical and economic indicators. Through the sintering experiments, we analyze the influence of material properties on the sintering process parameters and sinter yield and quality indexes, to establish the input-output relationship of prediction model. Then we built the model with support vector machine, that the accuracy rate reached more than83%. The system software was developed by C#programming language; its main function includes materials database management, iron ores proportion optimization and sintering burden calculation. Through practical application shows that the system can give an iron ores blending scheme that has better technical and economic indicators, based on the user demand and raw material conditions. |
PDF Full Text Request