电力信息与通信技术2026，Vol.24Issue(3)：52-59,8.DOI:10.16543/j.2095-641x.electric.power.ict.2026.03.07
基于多尺度用电特征融合的低压双路电源同供隐患研判算法
An Algorithm for Identifying Hidden Dangers of Low-voltage Dual-power Supply Simultaneously Based on Multi-scale Electricity Consumption Feature Fusion
摘要
Abstract
Low-voltage dual-power supply simultaneously is a common abnormal power safety hazard on the low-voltage side.Due to the failure of the dual-power supply switch to be installed or malfunctioning,the two distribution transformers are directly connected on the low-voltage side.In serious cases,it may cause line fires,electric shock injuries and deaths.It is highly concealed and difficult to detect,and it is impossible to detect problems in a timely and comprehensive manner by relying solely on on-site inspections by operation and maintenance personnel.This paper proposes an intelligent algorithm for identifying hidden dangers of low-voltage dual-power supply simultaneously based on multi-scale electricity consumption features fusion.It extracts three key features of statistical scale,time scale and spatial scale from electricity consumption current,power factor and other data,and then builds a co-supply hidden danger identification model called DangerSense based on multi-scale electricity consumption features fusion.It uses electric power big data analysis and artificial intelligence algorithm to directly locate safety hazard points,transforming traditional manual door-to-door inspections into data-driven intelligent identification,greatly improving the efficiency of problem verification by operation and maintenance personnel.Through case analysis,the average recognition Fl score of DangerSense model proposed in this paper reaches 90.52％.Compared with multiple benchmark algorithms,the method proposed in this paper has achieved significant improvement and has strong practical value.关键词
低压双路电源同供/异常用电/多尺度用电特征融合/电力大数据分析/人工智能Key words
low-voltage dual-power supply simultaneously/abnormal electricity consumption/multi-scale electricity consumption features fusion/power big data analysis/artificial intelligence分类
信息技术与安全科学引用本文复制引用
周祉君,陈锦铭,蔡云峰,赵新冬,陈烨..基于多尺度用电特征融合的低压双路电源同供隐患研判算法[J].电力信息与通信技术,2026,24(3):52-59,8.基金项目
国网江苏省电力有限公司科技项目"数字电网精准映射与计算推演技术研究"(J2023121). （J2023121）
