多支持向量机在高炉喷煤规则挖掘中的应用

Application of Multiple Support Vector Machines Model for Coal Injection Rules Mining in Blast Furnace

高炉冶炼受生产过程运行规律和生产技术人员操作方式支配,由于对高炉的认识还不充分,在高炉冶炼操作过程中有时忽略了工艺指标与参数之间的相互关系,造成能源的消耗。寻求高炉冶炼喷煤过程的操作规则对生产节能降耗就显得尤为重要。选取蕴含有客观规律和生产技术人员操作决策模式等重要信息的生产历史数据,选出满足优化标准的数据,运用模糊C均值聚类将数据分成子集,分别建立相应的支持向量机子模型,将输入对应每一类的隶属度值再作为权值进行加权求和得到模型最终的输出值,从而建立起高炉冶炼喷煤的操作规则模型。通过规则挖掘实现对高炉现场冶炼过程的指导,达到优化高炉生产操作模式的目的。

The blast furnace smelting is controlled by the production process operation rules and production techni- cal personnel operating model, due to people＇s understanding of the blast furnace is not enough, the relationship between the technological indicators and parameters in the process of blast furnace smelting operation is ignored sometimes which result in the energy consumption. It is particularly important to find the operation rules in the process of pulverized coal injection for blast furnace for saving energy and reducing consumption. The production history data that involve important information, such as objective laws and the production technical personnel op- erating decision model, were selected, data meeting optimization standard were then chosen and divided into sub- sets by the means of the fuzzy C-means clustering, and then the corresponding support vector machine model （SVM） was established separately, the input value corresponding to each kind of membership was used for weigh- ted sum and the final output value was obtained accordingly, from which the operation rule model of coal injection for blast furnace smelting was established eventually. The guidance for the blast furnace smelting process by the rule mining was implemented, and the goal of optimization of blast furnace production operation mode was reached.