基于用户负荷的用电模式分析方法

Power Consumption Analysis Method Based on User Load

针对目前不同行业不同用电类别的专变用户存在多种多样的用电行为,在用户用电异常分析的过程中,无法准确地判定出当前用户的用电行为是否异常的现状,提出了一种基于用户日负荷数据运用数据挖掘算法而建立的用电模式异常识别方法。该方法以用户日瞬时量数据为研究对象,首先对数据进行预处理和归一化,引进改进后的K-means聚类算法构建单一用户历史用电模式;其次利用用户测试数据到簇中心点的距离大于该簇的阈值半径定义为异常用户。最终通过量化的概率指标输出用户是否为用电模式异常,以此作为稽查人员判定用户用电异常的依据和参考。利用该方法对某供电单位专变用户进行用电行为异常进行判定,结果表明该方法处理效率较高,而且异常判别简单快捷,具有很好的实用价值。

In view of the present situation that most power supply enterprise marketing audit mainly rely on passive methods such as artificial inspection,which are hard to detect abnormal electricity customers,an electricity price implementation online inspection model is developed in this paper based on data mining technology. Using the mass data of measurement automation system and marketing system,the model firstly uses K-means clustering algorithm to construct the typical electricity track module to identi fy the customer's typical electricity mode. Secondly,the Mahalanobis distance discriminant analysis algorithm is adopted to establish the abnormal customer distinguish module,which can identify abnormal electricity customers automatically. The outputs of the model are all the suspected electricity customers,which can provide power inspectors audit scopes and basis. The feasibility and validity of the proposed method are verified by analysis results of the power marketing inspection work in one regional power supply bureau in South China.