流程工业中的数据挖掘建模与应用

Modeling of Data Mining and its Application in Process Industry

数据挖掘是应用一系列技术从大型数据库或数据仓库中提取人们感兴趣的信息和知识,已在零售业得到了广泛的应用,却在流程过程中使用较少。针对某石化企业的连续重整过程,利用决策树归纳法,建立反应器温度波动模型,对实时数据进行预测,对于引起反应器温度波动较大的操作提出建议性的决策树规则,从而达到改善控制效果和提高装置运行平稳性的目的。

Data mining （DM） is the process of extracting desirable knowledge or patterns from existing databases or dataware house for speci？c purposes, Much research has been done focusing on its application to retail sales, but it is seldom applied in process industry. Based on the characteristic of Continuous Catalyst Regeneration （CCR） process and decision tree induction, a reactor temperature fluctuation model is established to predict the temperature fluctuation for the current operation. Using the generated decision rules, we can adjust the control function and improve the stability of the devices.