Study On Optimization Model Of Iron-making Sintering Blending And Its Application

On the premise of guaranteeing product quality,how to reduce sinter cost by optimizing the rational mixing of various raw materials has always been an important topic for iron and steel enterprises.At present,the majority of domestic iron and steel enterprises mainly set up sintering blending schemes through the experience accumulated by the foremen in the first line of production for many years.Because of the subjective difference of artificial experience,it is difficult to achieve the optimal cost and sinter quality in the adjustment of ore blending scheme.Moreover,in the actual industrial production,the raw material composition changes frequently in the daily ore blending scheme,which increases the complexity of ore blending.Sintering blending is essentially a goal optimization problem with non-linear constraints.Based on the current optimization theory,the mathematical model is established.Based on the fast processing ability of computer,the optimal production plan can be obtained quickly,accurately and efficiently,and used in the control of production equipment to obtain maximum production benefits.The main research work of this paper is as follows:(1)A sinter blending model aiming at the optimal cost was designed.Catastrophe operation is added to genetic algorithm,and catastrophe effect is used to improve the global search performance of population in the process of genetic operation.Based on the principle of adaptive adjustment of crossover probability and mutation probability with evolutionary algebra in adaptive genetic algorithm,crossover operation and mutation operation are adaptively adjusted with the catastrophe cycle.After the catastrophe is completed,the individual mutation ability is improved and the local optimization ability of the algorithm is enhanced.By comparing with the actual data of nearly 100 batches of production in a steel plant,the sinter cost can be reduced by 3-10 yuan/ton under the condition of meeting the requirements of sinter physical and chemical indexes of enterprises.(2)A multi-objective sintering blending model with the best cost and the lowest sulfur content was designed.By discussing the problems in the multi-objective optimization process,a new fast non-dominated sorting genetic algorithm(NSGA-II)is proposed,which reduces the problem of too high individual repetition rate by using de-duplication operation,and reduces the influence of extreme target value individuals on the frontier solution set by adding extreme filter operation in the fast non-dominated layer sorting.After the ranking of non-dominant solutions,the individuals in the same non-dominant layer are clustered to reduce the number of individuals in the dense region and fill the sparse region,whichenhances the uniformity of individual distribution in Pareto decomposition and ensures the continuity of the knowledge set.Compared with the actual production data,the cost per ton of Pareto's concentrating ore blending scheme decreases while the sulfur content decreases by about 10% on average.(3)Using C # language,the research results of this paper are realized in engineering,and the "optimization calculation system of sinter blending" is developed.The application results of cooperative enterprises show that the overall performance of the research results satisfies the production needs of enterprises and achieves the expected results.It plays an important role in reducing sinter cost and improving sinter quality,and receives good social and economic benefits.