基于聚类分群的线损特征分析方法

电力用户窃漏电和异常用电的重要评价指标包括：电量异常、负荷异常、终端报警、线损异常等，其中线损异常是最为直接和最重要的指标．基于数据挖掘相关技术，提出一种线损分析方法；通过提取分线线损的表征变量，建立线损分析的指标体系；并应用K—means聚类算法对特征样本进行聚类分群，基于线损聚类分群类别分析线损特征规则；最后经实例分析证明了该方法的有效性．

Cluster algorithm based line losses analysis method

The important analysis indexes for abnormal electricity consumption are electric power, load, terminal alarm and line losses etc, wherein the line losses is the most critical and important index. Using the technology of data mining, a new method for line losses analysis was proposed in this paper. The influence indexes for line-loss analysis was built, which gave the detailed cal- culation process for the index parameters. And the K-means algorithm of data-mining was used to classify the line-loss samples. Then every classification group based line losses characteristics were analyzed in detail. Cases analysis results showed that the proposed method was highly effective.