高维电力数据的聚类优化算法的研究

Analysis of Clustering Method of Load Curve in Power System

为了加强智能电网构建的推进,加强智能量测终端大量投入使用,提出了聚类方法用于电力公司实时分析负荷特性及负荷用电数据信息。在分析电力负荷曲线数据的基础上,分别考虑了电网企业和社会价值信息,提出了基于FCM和K-means聚类2种方法来提取出用户用电规律,并分析用户负荷构成与用电行为。经过仿真分析和实验验证,通过均方误差、均值适宜度、聚类离散度、聚类时间等多个评判聚类效果好坏的指标进行聚类算法性能评价。这对于用户了解用电规律用电分配需求、指导电价制定具有参考价值。

In order to strengthen the advancement of smart grid construction and strengthen the large-scale use of smart measurement terminals,a clustering method is proposed for real-time analysis of load characteristics and load power data information of power companies.Based on the analysis of power load curve data,the power grid enterprises and social value information were considered separately,and two methods based on FCM and K-means clustering were proposed to extract user power consumption rules,and analyze user load composition and power consumption behavior.After simulation analysis and experimental verification,the performance of the clustering algorithm is evaluated by multiple indicators such as mean square error,mean suitability,cluster dispersion,and clustering time.Deeply grasp the rules of electricity consumption by users,which provides a reference for him to distribute electricity demand and guide electricity pricing.