摘要
提出了采用模糊聚类的方法,采集7种易于获得的高炉操作数据进行聚类分析,再结合神经网络训练,自行生成初始知识库,最终形成基本炉况判断专家系统。以随机采取的某445 m3高炉的现场数据(共150组)进行了验证性分析,结果表明,聚类方法适当,判断分析结果与高炉操作者的判断结果高度吻合,说明这是一种可行的高炉智能控制的新方法。
A method using fuzzy clustering to build BF expert system was studied. Seven easy to-handle operational variables were used to cluster at first. Artificial neural network was used to train data after cluster analysis. The initial knowledge base was self created. On the basis of knowledge base, an expert system for judging BF operation was generated at last, 150 groups of production data were used to test the system. The results conformed to the judgment of operators.
出处
《钢铁研究学报》
CAS
CSCD
北大核心
2006年第9期56-58,共3页
Journal of Iron and Steel Research
基金
2002年度重庆市科技计划资助项目(7263)
关键词
高炉冶炼
人工智能
模糊聚类
BF ironmaking
artificial intelligence
fuzzy clustering
