基于负荷控制潜力量化模型的工业用户群体画像方法  

Industrial customer group portrait method based on potential quantization model of load control

作　　者：陈逸涵 李扬[1] 沈运帷 CHEN Yihan;LI Yang;SHEN Yunwei(School of Electrical Engineering,Southeast University,Nanjing 210096,China)

机构地区：[1]东南大学电气工程学院,江苏南京210096

出　　处：《电力自动化设备》2021年第8期208-216,共9页Electric Power Automation Equipment

基　　金：国家电网公司科技项目(52182019000J)。

摘　　要：在中国新一轮电力体制改革背景下,研究工业用户负荷参与负荷控制的潜力,对于促进电网安全稳定运行有着积极的作用。以工业用户负荷数据为基础,充分提取用户在不同时间量度下的用电特征,从错时潜力、轮休潜力和避峰潜力3个方面构建多时间尺度负荷控制潜力指标体系。进而将信息熵和逼近理想解排序法相结合,构建负荷控制潜力量化模型,实现了对负荷控制潜力价值的衡量。并且,利用近邻传播算法分析量化模型结果,将用户按照潜力量化值聚类,实现了对不同用户特征群的划分。最后,基于某地区的实际工业用户负荷数据进行算例分析,实现了对工业用户群体负荷控制潜力画像结果的呈现。Under the background of the power system reform in China,research on exploiting the potential of industrial customer load to participate in load control is necessary to enable the safe and stable opera⁃tion of power grid.The electricity consumption characteristics under different time measurements are fully extracted based on the load data of industrial customers.A multi-time scale load control potential index system is constructed from three aspects:off-time potential,rest potential and peak avoidance potential.Then,the potential quantization model of load control is developed to provide the value assessment of control potential by combining information entropy and TOPSIS(Technique for Order Preference by Similarity to Ideal Solution).Moreover,the affinity propagation algorithm is employed to analyze the quantization model results,and the customers are clustered according to the potential quantization value to provide the classifica-tion of different customer characteristic groups.Finally,an example based on the actual load data of indus⁃trial customer in a certain region is conducted to present the load control potential quantization portrait results of industrial customer group.

关 键 词：负荷控制 潜力评估 工业用户 量化模型 近邻传播算法 用户画像 

分 类 号：TM714[电气工程—电力系统及自动化]