大用户多维度可中断特性精细化分析

Refined Analysis of Large-Consumers’ Interruptible Features from Multi-Dimension

大用户可中断特性的研究是实施可中断负荷的基础。该文选用Ward系统聚类法与考虑邻近相关度的改进模糊C均值聚类法对历史日负荷数据聚类分析,并利用有效性函数选择最佳聚类数目;提出二次分类模型,通过两步聚类得到精细化的负荷数据分类,提炼每一类聚类的中心曲线作为大用户典型用电模式;比较几类典型负荷曲线,从中断调控方式、中断容量、中断时长及中断时刻4个维度分析大用户的可中断特性;最后基于实际数据成功挖掘了棉纺印染加工行业某一用户的多维度可中断特性,验证了该文所提方法及模型的有效性。

The study of large-consumers’interruptible features is the basis to schedule interruptible loads.This paper chose the Ward system clustering method combined with improved fuzzy C-means clustering method(FCM)which takes into consideration the proximity correlation of each sample to deal with the consumers’daily load data.The number of clusters was optimized by the validity function.Then the twice classification model was proposed to analyze large users’interruptible characteristics.The refined load data classification was obtained by two-level clustering and the cluster centers were extracted as typical power consumption patterns.By comparing the several typical load curves,the interruptible characteristics of large users were analyzed from four dimensions:interruptible schedule mode,interruptible capacity,interruptible span and interruptible time.Finally,based on the actual historical data,the multi-dimensional interruptible characteristics of a user in the cotton textile printing and dyeing industry were successfully excavated,which verifies the effectiveness of the proposed method and model.