摘要
文中结合智能电能表计量大数据,对一定区域内的负载进行非侵入式多节点短期预测的方法进行了研究。同时研究了集聚效应对负载建模和预测的影响,说明了多节点预测的优势与必要性;评估了影响电能需求的变量,并对数据集进行特征选择。使用多元线性回归模型对自顶向下与自底向上两类预测方法进行了比较,在真实数据上的测试表明,以智能电能表计量大数据为支撑的自底向上方法在短期多节点负载预测上具有优势。
In this paper,a non-intrusive short-term multi-nodal load forecasting method in a certain area is proposed through combining with big data collected by smart meters.The aggregation effects on modelling and forecasting of load are studied,indicating the supremacy and necessity of multi-nodal forecasting.The variables affecting electricity demand are evaluated,and the feature selection of dataset is also carried out.Top-down and bottom-up forecasting approaches are compared through the multiple linear regression models.Experimental results on real-world dataset show that the bottom-up approach supported by measurement big data of smart meter has the advantage in short-term multi-nodal load forecasting task.
作者
王智
陈福胜
胡军华
杨静
苏玉萍
Wang Zhi;Chen Fusheng;Hu Junhua;Yang Jing;Su Yuping(State Grid Hunan Electric Power Co.,Ltd.,Changsha 410004,China)
出处
《电测与仪表》
北大核心
2022年第4期29-33,共5页
Electrical Measurement & Instrumentation
基金
国家重点研发计划项目(2018YFF0212906)。
