基于深度学习的输电线路故障类型辨识-中国电机工程学报.PDF

第39 卷 第1 期 中 国 电 机 工 程 学 报 Vol.39 No.1 Jan. 5, 2019 2019 年1 月5 日 Proceedings of the CSEE ©2019 Chin.Soc.for Elec.Eng. 65 DOI ：10.13334/j.0258-8013.pcsee.181024 文章编号：0258-8013 (2019) 01-0065-10 中图分类号：TM 726 基于深度学习的输电线路故障类型辨识 1 1 2 1 1 1 徐舒玮 ，邱才明 ，张东霞 ，贺兴 ，储磊 ，杨浩森 (1．上海交通大学大数据工程技术与人工智能研究中心，上海市 闵行区 200240 ； 2．中国电力科学研究院有限公司，北京市 海淀区 100192) A Deep Learning Approach for Fault Type Identification of Transmission Line 1 1 2 1 1 1 XU Shuwei , QIU Caiming , ZHANG Dongxia , HE Xing , CHU Lei , YANG Haosen (1. Research Center for Big Data and Artificial Intelligence Engineering and Technologies, Shanghai Jiaotong University, Minhang District, Shanghai 200240, China; 2. China Electric Power Research Institute, Haidian District, Beijing 100192, China) ABSTRACT: With new energy access and the emergence of 统为例，对所提方法进行验证。仿真结果表明，文中所述方 stochastic load behavior, feature construction and selection for 法可基本达到 99.99%的辨识正确率，不受故障线路、故障 transmission line faults have become increasingly difficult. On 阻抗、故障位置、电压扰动、频率扰动和负荷波动的影响， the basis of deep learning theory, a novel method of fault type 且能够应对电网运行过程中的噪声干扰。 identification for transmission line was proposed to realize 关键词：故障辨识；深度学习；PSS\E ；Ornstein-Uhlenbeck automatic characteristic learning, and Ornstein-Uhlenbeck 过程 process was introduced to better mimic electricity load fluctuation in stochastic distributed power grid. Firstly, a 0 引言 massive fault dataset for