摘要
随着电力行业市场化改革的深入,对用电客户进行细分并提供差异化服务已经成为必然趋势。针对电力用户的用电负荷数据特点,提出了一种基于大数据技术的用电特征相似性挖掘方法。采用DTW算法对负荷曲线相似度进行度量,并利用K-means算法对DTW距离矩阵进行聚类分析,实现用户负荷曲线的聚类和负荷特性分析。最后,以纺织印染业大工业用户的负荷数据为例进行验证,结果表明,该算法组合能够较好地反映负荷曲线的相似度,负荷曲线特征呈现显著差异。
With the deepening of market-oriented reform of power industry, it is a trend to focus on power user subdivision and provide customers with differentiated services. According to the characteristics of electrical load data, the paper proposes a big data-based method for similarity mining of power consumption feature.The similarity among load curves is calculated by DTW method, and the K-means clustering algorithm is employed for cluster analysis on DTW distance matrix to cluster load curve of users and analyze load characteristics. Experiments on the load data of large textile printing and dyeing industry users show that the proposed method can well reflect the similarity among load curves, and significant difference is shown in load curve features.
出处
《浙江电力》
2017年第12期37-41,共5页
Zhejiang Electric Power
关键词
售电市场
大数据技术
用电特征分类
动态时间规整算法
K-MEANS算法
power sales market
big data technology
means clustering algorithmpower consumption feature classification
DTW
Kmeans clustering algorithm