基于云计算的电能计量数据异常分析方法

An Analysis Method of Electric Energy Metering Data Abnormal Based on Cloud Computing

由于原始电能计量数据受到外界干扰,对数据实际异常的分析结果存在一定误差,为此,提出基于云计算的电能计量数据异常分析方法。利用云计算挖掘有效数据的特征,并以其为基础修复缺陷数据,对于重复数据直接删除,其余类型的脏数据以时间为基准进行标记,在数据分析阶段按照非目标数据源处理,利用加权回归的方式得出处理后数据的异常情况。测试结果表明,设计的方法可以实现对单一随机异常、多个随机异常以及多个连续异常的准确分析,并未出现异常分析丢失或异常分析错误的情况。

Because the original electric energy measurement data is interfered by the outside world,there is a certain error in the analysis results of the actual abnormal data.For this reason,a cloud computing-based abnormal analysis method of the electric energy measurement data is proposed.Use cloud computing to mine the characteristics of valid data,and repair defective data based on it,delete duplicate data directly,and mark other types of dirty data based on time.In the data analysis stage,it is processed according to non-target data sources,and weighted regression is used.The test results show that the design method can realize the accurate analysis of a single random anomaly,multiple random anomalies and multiple continuous anomalies,and there is no abnormal analysis loss or abnormal analysis error.