基于关联分析和梯度提升决策树的低压接线错误漏电用户定位
CSTR:
作者:
作者单位:

(国网湖南省电力有限公司长沙供电局,湖南 长沙 410004)

作者简介:

通讯作者:

陈超强(1973—),男,硕士,高级工程师,主要从事配网运维管理研究;E?mail:chencq@hn.sgcc.com.cn

中图分类号:

TM863

基金项目:

国网湖南电力公司科技项目(SGHNCS00PDJS2311137)


Localization of leakage caused by low‑voltage wiring errors on user side based on correlation analysis and gradient boosting decision trees
Author:
Affiliation:

(Changsha Power Supply Bureau, State Grid Hunan Electric Power Corporation, Changsha 410004, China)

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    用户侧中性线、地线接线错误在低压台区大量存在,使得用户负荷电流转化为剩余电流、导致台区漏电保护频繁跳闸并被迫退出运行,对用电安全构成威胁。接线错误隐藏在用户内部,难以定位排查,是台区漏保投运困难的重要原因。利用接线错误用户的负荷电流和台区剩余电流之间存在显著关联的特点,提出了基于关联分析和梯度提升决策树(gradient boosting decision tree,GBDT)的接线错误漏电用户定位方法。首先,对台区剩余电流和用户负荷电流的关联性进行定性和定量分析,基于皮尔逊相关系数判断其是否存在因果关联;然后,构建各用户负荷电流与台区异常剩余电流的GBDT模型,计算各用户的重要性评分大小,以衡量各用户对台区剩余电流异动的贡献程度;最后,进一步精准识别异常用户。实验结果表明,所提方法在复杂故障场景下也具有精准的异常用户识别能力。

    Abstract:

    Neutral and ground wiring errors on the user side are common in low-voltage transformer districts, which can cause user load currents to be converted into residual currents, resulting in frequent tripping of residual current circuit protection devices in the transformer district and their forced deactivation, thereby posing a threat to electrical safety. Since such wiring errors are present on the user side and are difficult to locate and troubleshoot, they have become a major obstacle to the deployment of residual current protection in transformer districts. This paper proposes a method for locating the user-side leakage due to wiring errors based on the significant correlation between user load currents and the residual current in the transformer districts, using correlation analysis and a gradient boosting decision tree (GBDT). The method starts with a qualitative and quantitative analysis conducted on the correlation between the residual current in the transformer district and the load current of users, with the Pearson correlation coefficient used to determine whether a causal relationship exists. A GBDT model is then constructed for the load current of each user in relation to the abnormal residual current in the transformer district, and importance scores are calculated to measure each user’s contribution to the fluctuations in residual current. This allows for the precise identification of user-side abnormalities. Experimental results demonstrate that the proposed method can accurately identify user-side abnormalities even in complex fault scenarios.

    参考文献
    相似文献
    引证文献
引用本文

郑峻峰,陈超强,陈 凤,等.基于关联分析和梯度提升决策树的低压接线错误漏电用户定位[J].电力科学与技术学报,2025,40(3):104-113.
ZHENG Junfeng, CHEN Chaoqiang, CHEN Feng, et al. Localization of leakage caused by low‑voltage wiring errors on user side based on correlation analysis and gradient boosting decision trees[J]. Journal of Electric Power Science and Technology,2025,40(3):104-113.

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2025-07-29
  • 出版日期:
文章二维码