电器与能效管理技术 ›› 2020, Vol. 593 ›› Issue (8): 52-57.doi: 10.16628/j.cnki.2095-8188.2020.08.010

• 配网技术与系统 • 上一篇    下一篇

基于随机森林算法的配电线路短路故障分类*

刘树鑫1, 卓裕1, 李津2, 田二胜2, 刘洋2, 冯毅3   

  1. 1.沈阳工业大学 电气工程学院,辽宁 沈阳 110870
    2.国家电网许继集团有限公司,河南 许昌 461000
    3.国网邢台供电公司,河北 邢台 054001
  • 收稿日期:2020-05-20 出版日期:2020-08-30 发布日期:2020-09-10
  • 作者简介:刘树鑫(1982—),男,副教授,博士,研究方向为电器新技术应用。|卓 裕(1994—),男,硕士研究生,研究方面为配电网故障诊断。|李 津(1975—),男,高级工程师,主要从事电力营销工作。
  • 基金资助:
    * 国家自然科学基金(51977132);国网河北省电力有限公司科技项目(5204XQ190003)

Classification of Short Circuit Faults in Distribution Lines Based on Random Forest Algorithm

LIU Shuxin1, ZHUO Yu1, LI Jin2, TIAN Ersheng2, LIU Yang2, FENG Yi3   

  1. 1. School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,China
    2. State Grid XJ Group Corporation,Xuchang 461000,China
    3. State Grid Xingtai Power Supply Company,Xingtai 054001,China
  • Received:2020-05-20 Online:2020-08-30 Published:2020-09-10

摘要:

为实现对配电线路短路故障后的故障类型进行准确分类,提出一种结合随机森林算法分类器的故障分类方法。针对谐振接地系统下故障电流的特点,提出通过集成经验模态分解(EEMD)提取故障线路电流的暂态高频分量,建立用于区分不同故障类型的7个故障特征量。在PSCAD/EMTDC中建立仿真模型,通过改变过渡电阻、故障初始角、故障线路获得大量样本,对随机森林分类器进行训练和测试。测试结果表明所提方法具有高分类准确率。

关键词: 配电线路, 短路故障, 故障分类, 高频分量, 集成经验模态分解

Abstract:

In order to accurately classify the fault types after short-circuit faults on distribution lines,this paper proposed a fault classification method combined with a random forest algorithm classifier.According to the characteristics of the fault current in the resonant grounding system,an integrated empirical mode decomposition (EEMD) is used to extract the transient high-frequency components of the fault line current,and seven fault characteristic quantities are established to distinguish different fault types.A simulation model was established in PSCAD/EMTDC,and a large number of samples were obtained by changing the transition resistance,the initial angle of the fault,and the fault line.The random forest classifier was successively trained and tested.The test results show that the method has high classification accuracy.

Key words: distribution lines, short-circuit faults, fault classification, high-frequency components, ensemble empirical mode decomposition(EEMD)

中图分类号: