基于TwoStep算法的国丰1号高炉操作炉型聚类分析与应用  

Clustering Analysis and Application of Operative Profile in Guofeng No.1 Blast Furnace Based on the Algorithm of TwoStep

作　　者：曹英杰[1] 张建良[1,2] 国宏伟[1,2] 杜申[1] 白俊丽[1] 

机构地区：[1]北京科技大学冶金与生态工程学院,北京100083 [2]北京科技大学钢铁冶金新技术国家重点实验室,北京100083

出　　处：《钢铁》2013年第10期17-22,35,共7页Iron and Steel

基　　金：国家自然科学基金委员会与宝钢集团有限公司联合资助项目(51134008);国家自然科学基金资助项目(51204013);中央高校青年人才培养基金资助项目(FRF-TP-12-020A)

摘　　要：以国丰1号高炉为平台,将数据挖掘技术运用到操作炉型管理中,采用数据挖掘中的TwoStep算法,对高炉炉身冷却壁热电偶温度值进行聚类分析,研究了操作炉型变化与高炉生产指标之间的对应关系,总结出与透气性指数相关的炉型变化的部分规律,并以国丰1号高炉2012年11月实际生产数据为例,对聚类分析结果应用情况进行在线监测。实践证明,聚类分析结果可以有效监控炉型变化,对炉况监测提供了可靠信息。As the technology of data mining applied to the management of operation profile, making clustering analy sis of the thermocouple temperature of blast furnace cooling stave with the algorithm of TwoStep in data mining in Guofeng No. 1 blast furnace, the relationship between the change of operative profile and production index of blast furnace was built, and some rules on the change of operative profile related to gas permeability index were summa rized. The data on the process of practical production in Guofeng in November 2012 was taken an example to monitor the application results of the clustering analysis online. The practice has proved that the result of clustering analysis can effectively monitor and control the change of operative, and it can provide reliable information on the blast furnace monitoring.

关 键 词：操作炉型 TwoStep算法 聚类分析 透气性指数 在线监测 

分 类 号：TF57[冶金工程—钢铁冶金]