摘要
配变日负荷曲线可用于准确监测配变运行状态,文章采用特征工程与相似度聚类的方法生成负荷相似日簇,在此基础上使用集成强化学习构建了配变日负荷预测模型。经应用验证,该模型可对各配变负荷曲线进行准确预测,实现配变负荷状态预感知,为电网调度管理、运行检查工作提供支撑。
The daily load curve of distribution transformer can be used to accurately monitor the distribution transformer running status.In this paper,the method of feature engineering and similarity clustering is adopted to generate load similar daily clusters.On this basis,the enhanced integratedlearning is used to build a daily load prediction model of the distribution transformer.The application has verified that the model can accurately predict the power load curve of each distribution transformer and realize fore-perception of the load state,which provides support for power dispatching management and operation inspection.
作者
陈朔
王尉
章柯
王艳龙
张照
陈雪圆
CHEN Shuo;WANG Wei;ZHANG Ke;WANG Yanlong;ZHANG Zhao;CHEN Xueyuan(State Grid Hefei Power Supply Company,Hefei 230022,China;State Grid Feixi Power Supply Company,Hefei 231200,China)
出处
《安徽电气工程职业技术学院学报》
2020年第4期64-68,共5页
Journal of Anhui Electrical Engineering Professional Technique College
关键词
负荷曲线预测
负荷相似日
集成强化学习
load curve prediction
load similarity day
enhanced integratedlearning
