基于数据挖掘的电能量数据异常特征提取方法  

Abnormal feature extraction method of electric energy data based on data mining

作　　者：代庆 陈耀冲 张霞 DAI Qing;CHEN Yaochong;ZHANG Xia(Digital Grid Research Institute Co.,Ltd.,China Southern Power Grid,Guangzhou 510520,China)

机构地区：[1]南方电网数字电网研究院有限公司,广东广州510520

出　　处：《电子设计工程》2023年第1期129-132,共4页Electronic Design Engineering

基　　金：南网数研院电能量数据挖掘产品研发项目(670000HK42200011)。

摘　　要：以高精度、高效率提取电能量数据异常特征为目的,研究基于数据挖掘的电能量数据异常特征提取方法。基于改进快速密度峰值聚类算法,检测电能量数据样本集中异常数据,进行异常电能量数据特征分类,通过属性特征密度分类异常电能量数据特征,以优化后的参数为输入,获取异常电能量数据特征分类的更新方案,提取电能量数据异常特征。实验结果表明,所提方法对正向有功总电量、反向无功总电量、四象限无功电量三种电能量数据中的异常数据检测率最高,误报率最小,可提升电能量数据异常特征提取效果。In order to extract the abnormal features of electric energy data with high accuracy and efficiency,the method of extracting abnormal features of electric energy data based on data mining is studied.Based on the improved fast density peak clustering algorithm,the abnormal data in the electric energy data sample set is detected,and the abnormal electric energy data features are classified.The abnormal electric energy data features are classified by the attribute feature density.With the optimized parameters as the input,the update scheme of the abnormal electric energy data feature classification is obtained,and the abnormal electric energy data features are extracted.The experimental results show that the proposed method has the highest detection rate of abnormal data and the lowest false alarm rate among the three kinds of electric energy data,including forward active power,reverse reactive power and four quadrant reactive power,which can improve the effect of abnormal feature extraction of electric energy data.

关 键 词：数据挖掘 电能量 异常特征 数据提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]