摘要
针对烧结矿生产配料结构复杂多元,既要保证质量,又要实现效益最优的问题,依托于唐钢不锈钢公司自动控制系统中存储下来的海量历史数据,应用数据挖掘技术中的聚类算法对历史的配矿数据进行聚类分析,最终建立相应的神经网络模型,对新配矿数据按照烧结矿技术指标进行预测,从而指导并调整配矿比例。运行结果表明,该算法简单、结果准确。
To guarantee quality and realize optimized profit under the condition of complicated multi-element structure of ore proportioning in sinter production,based on the saved macro history data in automation control system of Stainless Steel Company of Tang Steel,the historical proportioning data is analyzed with clustering algorithm of data mining technology and at last an appropriate neural network model is constructed.With that the new proportioning data is predicted according to the technical index and used to guide and adjust proportioning ratio.It is showed from operation that the algorithm is simple and accurate.
出处
《河北冶金》
2012年第4期12-14,共3页
Hebei Metallurgy
关键词
数据挖掘
聚类
神经网络
烧结配矿
预测
data mining
clustering
neural network
sintering ore proportioning
predict
