基于DBSCAN算法的售电均价异常识别模型构建与应用  

Construction and application of the identification model of average electricity selling price anomaly based on the DBSCAN algorithm

作　　者：杨玉强 胡若云 YANG Yuqiang;HU Ruoyun(State Grid Zhejiang Electric Power Co.,Ltd.,Hangzhou 310007,China)

机构地区：[1]国网浙江省电力有限公司,杭州310007

出　　处：《浙江电力》2023年第4期72-78,共7页Zhejiang Electric Power

摘　　要：随着电力市场化改革推进,电力改革出现市场主体多元化、利益诉求多样化的新趋势。为了解决电力市场售电均价水平偏高,部分用户用电成本上涨的问题,以售电公司售电均价异常智能分析为切入点,构建售电公司代理零售用户市场化均价异常分析模型;运用DBSCAN聚类分析算法定位电价异常的零售用户,智能划分不同异常用户群体并进行溯源分析;分析结果便于电网公司主动服务用户降低成本,同时辅助政府决策,推动电力市场化改革。With the advancement of the market-oriented reform of electricity,a new trend of pluralistic market enti⁃ties and diversified interest demands has emerged in the power reform.In order to solve the problem of a high aver⁃age price level of electricity sold in the electricity market and the rising cost of electricity for some customers,the in⁃telligent analysis of average electricity selling price anomaly of electricity selling companies is conducted first,and an analysis model for market-oriented average electricity price anomaly of retail users with electricity sales company as their agent is built.The DBSCAN clustering algorithm is used to locate retail users with electricity price anomaly,intelligently divide different groups of customers with the anomaly and conduct traceability analysis so that power grid enterprises can proactively serve customers,reduce costs,assist government decision-making,and promote market-oriented reform of electricity.

关 键 词：零售均价 DBSCAN算法 异常挖掘 聚类分析 

分 类 号：TM73[电气工程—电力系统及自动化] F426.61[经济管理—产业经济]