湖南电力 ›› 2026, Vol. 46 ›› Issue (4): 106-112.doi: 10.3969/j.issn.1008-0198.2026.04.014

• 配电网与用能技术 • 上一篇    下一篇

基于居民电能表电流的漏电识别方法

王绍槐, 尹有鹏, 翁薇, 蒋卓宸, 尹一江   

  1. 国网湖南省电力有限公司株洲供电分公司,湖南 株洲 412000
  • 收稿日期:2026-04-14 修回日期:2026-05-19 出版日期:2026-08-25 发布日期:2026-09-11
  • 通信作者: 王绍槐(1967),男,研究员级高级工程师,研究方向为用电信息采集。
  • 作者简介:尹有鹏(2000),男,硕士研究生,研究方向为电力计量。翁薇(1984),女,高级工程师,研究方向为电力智能监测。蒋卓宸(1999),女,助理工程师,研究方向为电力智能监测。尹一江(1992),女,高级工程师,研究方向为电力智能监测。
  • 基金资助:
    国网湖南省电力有限公司职工技术创新项目(B716B025Z002)

Leakage Identification Method Based on Neutral and Live Line Currents in Residential Electricity Meters

WANG Shaohuai, YIN Youpeng, WENG Wei, JIANG Zhuochen, YIN Yijiang   

  1. State Grid Zhuzhou Power Supply Company, Zhuzhou 412000, China
  • Received:2026-04-14 Revised:2026-05-19 Online:2026-08-25 Published:2026-09-11

摘要: 针对居民侧电能表表后线路固定漏电隐蔽性强、现场排查效率低,以及现有方法对台区总剩余电流监测依赖较强、现场应用受限的问题,提出一种基于居民电能表电流的固定漏电识别方法。该方法利用居民户表24点电流日数据,构造电流绝对值差序列来表征用户回路不平衡程度;引入相邻时刻差值比值及其对1的偏差均方根,构建差值一致性指标;在此基础上,建立极端不平衡判据、差值幅值判据和差值一致性判据相结合的分层研判策略;采用统计分析与ROC曲线优化相结合的方法确定关键阈值。现场应用结果表明,在4个供电所40户经核验存在漏电的居民用户样本中,成功识别37户,检出率达92.50%,所提方法无需新增监测设备或改造台区,具有较好的工程实用性与推广价值。

关键词: 居民漏电, 电能表电流, 剩余电流, 异常识别

Abstract: To address the challenges of the high concealment of fixed leakage currents on the consumer side of the meter, low efficiency in on-site troubleshooting, and the limitations of existing methods which rely heavily on monitoring the total residual current of the transformer district and face constraints in practical application, a method for identifying fixed leakage currents based on the neutral and live line currents of residential electricity meters is proposed. This method utilizes only the daily data of the 24-point neutral and live line currents from residential meters to construct a sequence of absolute differences between neutral and live line currents, thereby characterizing the degree of circuit imbalance. The ratio of the difference between adjacent time points and the root mean square deviation from 1 is introduced to construct a difference consistency index. On this basis, a hierarchical analysis strategy is established that combines extreme imbalance criteria, difference amplitude criteria, and difference consistency criteria, and a method combining statistical analysis and ROC curve optimization is adopted to determine key thresholds. Field application results indicate that, among a sample of 40 residential users verified to have leakage currents across four power supply stations, 37 households were successfully identified, achieving a detection rate of 92.50%. The proposed method requires no additional monitoring equipment or transformer substation modifications, demonstrating good engineering practicality and potential for widespread adoption.

Key words: residential electrical leakage, neutral and live wire currents, residual current, anomaly detection

中图分类号: