电器与能效管理技术 ›› 2024, Vol. 0 ›› Issue (5): 67-74.doi: 10.16628/j.cnki.2095-8188.2024.05.009

• 电气设计与探讨 • 上一篇    下一篇

基于RDBN深度学习算法的窃电监测系统设计

陈志谦1, 鲍光海1, 方艳东2   

  1. 1.福州大学 电气工程与自动化学院,福建 福州 350108
    2.浙江天正电气股份有限公司,浙江 乐清 528300
  • 收稿日期:2024-01-22 出版日期:2024-05-30 发布日期:2024-06-25
  • 作者简介:陈志谦(1996—),男,硕士研究生,研究方向为电气装备测试与在线监测技术。|鲍光海(1977—),男,教授,博士生导师,研究方向为电器及其系统智能化与故障诊断。|方艳东(1973—),男,工程师,主要从事低压电器检测和装配设备自动化开发设计。
  • 基金资助:
    福建省科技计划项目(2023H0007)

Design of Electricity Theft Monitoring System Based on RDBN Deep Learning Algorithm

CHEN Zhiqian1, BAO Guanghai1, FANG Yandong2   

  1. 1. College of Electrical Engineering and Automation,Fuzhou University, Fuzhou 350108, China
    2. Zhejiang Tianzheng Electric Co.,Ltd., Leqing 528300, China
  • Received:2024-01-22 Online:2024-05-30 Published:2024-06-25

摘要:

窃电行为不仅会造成电网非技术性损耗增加,而且可能因操作不当影响电网设备的运行安全和窃电者的人身安全。针对当前电网在窃电检测方面存在的稽查难度大、检测效率低等问题,设计了窃电监测系统。配套监测装置可灵活安装在供电线路上,使用电流互感器取能,实时采集线路电流,利用4G模块将数据传输至云服务器,在上位机软件中采用实值深度置信网络(RDBN)算法对数据进行分析。仿真和实验测试表明,RDBN算法对窃电状态的识别准确率达到98.15%,监测系统能实时获取并分析监测数据,标记可疑窃电线路,降低稽查难度,提高检测效率。

关键词: 非技术性损耗, 窃电检测, 电流互感器取能, 实值深度置信网络

Abstract:

Electricity theft not only leads to an increase in non-technical losses in the power grid,but also has the potential to impact the operational safety of power grid equipment and pose risks to the safety of individuals involved in power theft due to improper handling.The current power grid faces challenges in detecting electricity theft,including difficulty in inspections and low detection efficiency.In response to these issues,a power theft monitoring system is proposed.The associated monitoring device can be flexibly installed on power supply lines,utilizing a current transformer for power supply.It real-time collects current from the power line and transmits the data to a cloud server using a 4G module.The upper-level software employs the real-valued deep belief network (RDBN) algorithm for data analysis.Through simulation and experimental testing,the RDBN algorithm achieves a recognition accuracy of 98.15% for power theft states.The monitoring system can acquire and analyze monitoring data in real-time,mark suspicious power theft lines,reduce inspection difficulty,and improve detection efficiency.

Key words: non-technical losses, detecting electricity theft, current transformer for power supply, real-valued deep belief network (RDBN)

中图分类号: