中国电力 ›› 2021, Vol. 54 ›› Issue (2): 52-57,65.doi: 10.11930/j.issn.1004-9649.202006006

• 国家“十三五”智能电网重大专项专栏:(五)电力传感技术及应用专栏 • 上一篇    

基于特高频无线智能传感阵列的敞开式变电站放电定位方法

闫帅1, 李朋宇1, 王高洁1, 李强1, 吴凡2, 罗林根2   

  1. 1. 国网安徽省电力有限公司亳州供电公司,安徽 亳州 236800;
    2. 上海交通大学 电气工程系,上海 200240
  • 收稿日期:2020-06-01 修回日期:2020-12-12 发布日期:2021-02-06
  • 作者简介:闫帅(1989-),男,工程师,从事电网检修试验,E-mail:820206915@qq.com;吴凡(1994-),男,通信作者,硕士,从事电力设备状态评估研究,E-mail:wufan1005@qq.com;罗林根(1982-),男,博士,副研究员,从事电力设备状态评估研究,E-mail:llg523@sjtu.edu.cn
  • 基金资助:
    国家重点研发计划资助项目(2017YFB0902705)

Partial Discharge Localization Method Based on UHF Wireless Sensor Array in Air-insulated Substation

YAN Shuai1, LI Pengyu1, WANG Gaojie1, LI Qiang1, WU Fan2, LUO Lingen2   

  1. 1. State Grid Bozhou Power Supply Company, Bozhou 236800, China;
    2. Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
  • Received:2020-06-01 Revised:2020-12-12 Published:2021-02-06
  • Supported by:
    This work is supported by National Key Research and Development Program of China (No.2017YFB0902705)

摘要: 为了实现对电力设备绝缘劣化的全面监测和敞开式变电站的故障预警,提出了一种基于接收信号强度和高斯过程分类的变电站空间局部放电源定位方法。相较于现有的特高频时差法,提出的方法只需要特高频信号幅值强度的测量值,而不是对特高频信号进行高速同步采样,极大地降低了装置成本及改善了使用便捷性。同时,为了减轻特高频信号传输建模受到多路径效应和阴影效应的影响,提出了采用高斯过程分类算法用来学习不同传输路径下特高频信号特征,以生成决策区域来识别正确的接收信号强度数值,从而提高定位精度。现场试验结果表明,提出的空间局部放电源定位方法的平均误差为2.51 m,满足识别变电站内可能存在绝缘劣化的电力设备以达到故障预警的目的。

关键词: 局部放电, 敞开式变电站, 特高频, 故障预警, 接收信号强度, 高斯过程分类

Abstract: To achieve comprehensive insulation deterioration monitoring of power equipment and fault early-warning of air-insulated substations, we propose a data-driven partial discharge (PD) source localization methodology based on received signal strength indicator (RSSI) and Gaussian process classification (GPC). Compared to the existing ultra-high frequency (UHF) time-difference-based technique, the proposed method needs only to record the RSSI measurements, while needs not require time synchronization between UHF sensors, which provides a low-cost solution with high adaptability. Meanwhile, to mitigate the multipathing and shadowing effects in modeling the UHF signal attenuation, it is proposed to use the GPC algorithm to learn the UHF signal features under different transmission paths so as to generate decision-making regions for identifying the RSSI values that are consistent with the signal attenuation model, and as a result, the localization accuracy is improved. The field test is performed and the results show that the proposed method has a mean PD source localization error of 2.51 meters, which is sufficient to identify the power equipment with insulation deterioration in a substation for fault early-warning.

Key words: partial discharge, air-insulated substation, UHF, fault early warning, received signal strength indicator, Gaussian process classification