基于数据挖掘的电能计量装置故障在线诊断方法

An Online Fault Diagnosis Method for Electric Energy Metering Devices Based on Data Mining

提出了一种基于数据挖掘的电能计量装置故障在线诊断方法。根据电能计量装置的拓扑关系和数据挖掘算法在事务集中生成的关联规则,对事务数据库进行扫描,将数据库压缩成一棵事务树,并按照数据种类进行分组,在树的节点中保存所有事务的关联信息。调整故障数据关联使其能够符合关联规则。运用决策树方法进行故障诊断,计算信息增益。根据决策树划分属性信息增益最大原则,选择最优的划分属性对数据样本进行分类,得到最终的叶节点。确定最优划分属性选取后,建立决策树模型。将最优属性数据值代入决策树模型进行预测,从而得到模型预测的故障诊断结果。

This paper presents an online fault diagnosis method of electric energy metering device based on data mining.According to the topological relationship of electric energy metering device and the association rules generated by data mining algorithm in the transaction set,the transaction database is scanned,compressed into a transaction tree,and grouped according to data types,and the association information of all transactions is stored in the nodes of the tree.Adjust the fault data association to make it conform to the association rules.The decision tree method is used for fault diagnosis and information gain calculation.According to the principle of maximum gain of decision tree partition attribute information,the optimal partition attribute is selected to classify the data samples and get the final leaf nodes.After determining the optimal partition attribute selection,the decision tree model is established.The optimal attribute data values are substituted into the decision tree model for prediction,so that the fault diagnosis results predicted by the model can be obtained.