摘要
电力用户负荷曲线聚类分析是电力数据挖掘中的一个研究热点。负荷曲线聚类之前需对负荷曲线进行标准化处理,现有研究尚没有可以对不同标准化方法下的负荷曲线聚类结果进行评价的指标。提出了一种与标准化方法无关的电力负荷聚类评价指标,首次将近邻传播算法应用在负荷曲线聚类中,并给出了应用聚类结果的建议。算例结果表明:峰值标准化方法具有较好的聚类效果,相对于传统的负荷曲线聚类方法,近邻传播算法具有更好的聚类效果。
Electric load profile clustering is a research hotspot for data mining in power system. Load profiles need to be normalized before clustering. There is still no index to evaluate the clustering results of different normalization methods in the literature. An index to evaluate clustering results that is proposed. Affinity propagation algorithm is used in load profile clustering for the first time, and recommendations for application of the clustering results are given. Case study shows the normalization method using peak load is better than other methods,and the affinity propagation algorithm is better than traditional methods for load profile clustering.
引文
[1]宋亚奇,周国亮,朱永利.智能电网大数据处理技术现状与挑战[J].电网技术,2013,37(4):927-935.
[2]CHICCO B G,NAPOLI R,POSTULACHE P,et al.Customer Characterization Options for Improving the Tariff Offer[C].//IEEE Transactions on Power Systems.2010.
[3]CHEN C S,HWANG J C,HUANG C W.Application of load survey to proper tariff design[J].IEEE Trans on Power Systems,1997,12(4):1746-1751.
[4]黄宇腾,侯芳,周勤,等.一种面向需求侧管理的用户负荷形态组合分析方法[J].电力系统保护与控制,2013,41(13):20-25.
[5]赵莉,候兴哲,胡君,等.基于改进k-means算法的海量智能用电数据分析[J].电网技术,2014,38(10):2715-2720.
[6]PANAPAKIDIS I P,ALEXIADIS M C,PAPAGIANNISG K.Enhancing the clustering process in the category model load profiling[J].IET Generation,Transmission and Distribution,2015,9(9):655-665.
[7]宋易阳,李存斌,祁之强.基于云模型和模糊聚类的电力负荷模式提取方法[J].电网技术,2015,38(12):3378-3383.
[8]TSEKOURAS G J,HATZIARGYRIOU N D,DIALY-NAS E N.Two-stage pattern recognition of load curves for classification of electricity customers[J].IEEETrans on Power Systems,2007,22(3):1120-1128.
[9]CHICCO G,NAPOLI R,PIGLIONE F.Comparisons among clustering techniques for electricity customer classification[J].IEEE Trans on Power Systems,2006,21(2):933-940.
[10]PIAO M,SHON H S,LEE J Y,et al.Subspace srojection method based clustering analysis in load profiling[J].IEEE Trans on Power Systems,2014,29(6):2628-2635.
[11]ZHANG T,ZHANG G,LUJ,et al.A new index and classification approach for load pattern analysis of large electricity customers[J].IEEE Trans on Power Systems,27(27):153-160.
[12]BREIMANL.Random forests[J].Machine Learning,2001,45(1):5-32.
[13]赵岩,李磊,刘俊勇,等.上海电网需求侧负荷模式的组合识别模型[J].电网技术,2010,34(1):145-151.
[14]唐学用,万会江,叶航超,等.贵州统调电网典型日负荷特性分析与预测[J].中国电力,2015,48(9):24-30.
[15]FREY B J,DUECK D.Clustering by passing messages between data points[J].Science,2007,315(5814):972-976.