摘要
不同的用户在用电规律性方面所表现的特征具有较大的差异。需要根据其所表现出来的特征设计不同的负荷预测模型以提高预测的精度。根据短期负荷在负荷曲线形状、负荷波动特性、负荷敏感因素方面的差异,分别对应地采用稳定度、平均负荷率、负荷率标准差3个指标评价来刻画短期负荷的特性;针对所提取的短期负荷特征,提出一种基于密度中心快速辨识和K-Means的两阶段短期负荷聚类算法;根据所得到的聚类结果,对于每一类负荷进行数据分析,归纳每一类负荷具有的特性,搭建短期负荷的标准化定义模型。该模型可适用于系统负荷、母线负荷、大用户负荷等不同形式的负荷分析。
Short-term loads of different consumers have distinct load patterns.In order to improve the accuracy of load forecast,it is of great significance to design specialized forecasting model for each type of load based on their characteristics.Considering the differences of short-term loads in terms of their load curve shapes,fluctuation and sensitive factors,three indices are chosen to describe the features of short-term loads,namely,load stability,average load rate,and standard deviation.A two-step clustering algorithm that uses density centers and K-Means approaches is proposed to categorize the consumers according to their features.Analysis on clustering results of short-term load is carried out.The features of clustering results are summarized to form the standard definition model for short-term loads.The proposed model is suitable for the analysis for various kinds of loads,such as system load,bus load and large consumer load.
出处
《电力科学与技术学报》
CAS
北大核心
2016年第1期3-10,共8页
Journal of Electric Power Science And Technology
基金
国家杰出青年科学基金(51325702)
国家电网公司科技项目(SGHB0000KXJS1400044)
关键词
短期负荷
标准化定义模型
特征提取
两阶段聚类
short-term loads
standard definition model
feature extraction
two-stage clustering