摘要
为了实现配电网侧差异性规划和精益化管理需求,对单元制配电网展开精细化负荷预测已经成为新的发展趋势。首先利用改进K-means方法,根据历史负荷数据,对区域内的负荷进行分类,并匹配到单元制配网的实际负荷类型;然后结合负荷密度、用户数量、天气状况、国民经济等因素,应用回归方法归纳出各类型负荷的典型单位曲线,实现各类负荷的单独预测;最后考虑节假日和用电习惯等主客观因素对预测结果进行校正,叠加各类负荷实现单元制配网总负荷的预测。算例表明,基于大数据和多因素的单元制配网精细化负荷预测方法不仅能够提升配网负荷的预测精度,还能依据负荷类型划分结果指导配网规划和扩建。
In order to realize the differential planning and refined management of distribution network,the refined load forecasting of unit distribution network has become a development trend.Firstly,improved K-means method is used to classify the load in the area according to the historical data,and match it with the actual load type of the unit distribution network.Then,combined with the load density,the number of users,weather conditions,the national economy,etc.,regression method is used to obtain the typical unit curve of various kinds of load,and realize the individual prediction of various kinds of load.Finally,considering the factors such as holidays and electricity consumption behavior,the prediction results are corrected,and various kinds of load are superimposed to predict the total load of the unit distribution network.The case shows that the refined load forecasting method based on big data and multi-factors not only can improve the load forecasting accuracy of the distribution network,but also can guide the distribution network planning and expansion according to the load classification.
作者
李富鹏
沈秋英
王森
王承民
谢宁
LI Fupeng;SHEN Qiuying;WANG Sen;WANG Chengmin;XIE Ning(State Grid Suzhou Power Supply Company,Suzhou 215000,China;State Grid Suzhou Wujiang Power Supply Company,Suzhou 215200,China;School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《智慧电力》
北大核心
2020年第1期55-62,共8页
Smart Power
基金
国家重点研发计划资助项目（2018YFB1503000）~~
关键词
电力大数据
网格划分
单元制
负荷预测
big data for power
mesh generation
unit system
load forecasting
