常减压装置生产数据的聚类分析

为了研究数据挖掘技术在过程工业数据中的应用,采用基于Bayes原理的AutoClass聚类算法,对炼油厂常减压装置实际生产数据进行了聚类分析处理。结果表明,经过聚类分析可以从历史数据记录中提取清晰的类别信息,其中包含有用的知识。同时还通过对比实验,分析了多个聚类参数对结果的影响及原因。最后分析了聚类结果,讨论了应用该结果的方法。

Clustering analysis of atmospheric and vacuum distillation unit operational data

Data mining is a new subject that appeared recent years. This paper considers a data mining application based on clustering algorithm AutoClass, by a case study of operational data of an atmospheric and vacuum distillation unit. Processing these plant operational data by clustering analysis can get classification results clearly, which can provide valued information of operating state, optimization and modeling. The paper also analyzes the effects and reasons of varied classification parameters to clustering results.