The Research Of Optimal Model For Sintering Processing Parameters

With the constant change of global iron ore resources and the request of lower cost-higher profit in steel enterprises, the kinds of iron ores in sintering process are constantly adjusted. In order to attain the anticipatable sintering quality, the suitable sintering parameters must be provided fast and accurately when ore mix scheme is changed. However, in the past the adjustment of sintering parameters frequently depends on the historical experiences or the results of sintering trial when the ore resources are changed. And those repeated adjusting courses often take a long time. Therefore, the course of modeling for sintering process, the course of process parameters optimizing based on those sintering mathematics model and the course of rational parameters adjusting while mix scheme is changed become the critical point for sintering production and steady controls.With the artificial neural networks and genetic algorithms optimizing technology, the optimization model of the sintering process parameters is researched. Two kinds of optimization logic model of sintering process are put forward after analyzing the relation of mix ore scheme , process parameters and product quality index of Baoshan iron plant. And a large number of entity models are set up with the past five years sintering data. Eventually the optimization model of sintering parameters for Baoshan iron and steel plant is attained through comparing and analyzing those entity models.In the course of systematic modeling, the artificial neural networks method is studied and three-layer network structure is confirmed. Some improving measures for the defect of traditional back propagation network algorithms are studied, and the better application result is acquired.The artificial neural networks modeling platform is studied and developed. At the same time those problems such as module, flexible degree, transplanting of modeling and so on are solved as well. By adopting real number code programming, the genetic algorithms optimizing solving platform is developed. In the course of parametersoptimizing of models, the optimization method of sintering parameters-GA-BP, is proposed and realized. The optimization method of GA-BP supplies a kind of feasible mode for solving the problem of sintering parameters optimizing in Baoshan iron and steel plant.The development of optimizing and guiding systematic software of Baoshan iron and steel plant is realized by adopting the mix the programming strategy of MATLAB and VC++ and using ADO database technology. This software system is collected, parameter optimize, GA-BP optimize on an organic whole, and provides one good platform for solving the problem of sintering parameters optimizing in Baoshan iron and steel plant. The prediction rate of the model, which base on five years sintering data of Baoshan iron and steel plant, is more than 85%. The model can realize the optimization of process parameters and provide kinds of optional optimization scheme, which provides the software system flexibility and validity of practical application.