基于格拉姆角场-改进残差网络的小电流接地系统故障选线

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (2178 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对小电流系统发生单相接地故障时特征信息不明显、现有选线方法易受故障情况和环境噪声影响的问题,本文提出一种基于格拉姆角场（GAF）和改进残差网络（PrResNet）的小电流接地故障选线新方法。首先,搭建仿真模型采集零序电流等数据,利用GAF变换将一维零序电流时序信息转换成二维图像信息;然后,对残差网络（ResNet）进行改进,提出PrResNet用于提取图片特征;最后,通过Softmax输出选线结果。通过测试GAF-PrResNet在配电网不同运行工况和故障条件下的选线效果,证明了基于GAF-PrResNet的故障选线模型能够在强噪声干扰场景下实现不同故障位置和过渡电阻条件下的高精度选线,且具有较强的鲁棒性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	刘伟
	杨东风
	王洪志
	闫文迪

关键词 ：小电流接地配电网, 故障选线, 格拉姆角场（GAF）变换, 改进残差网络（PrResNet）

Abstract：In order to solve the problem that the characteristic information is not obvious when single-phase grounding fault occurs in a small current system, and the existing route selection method is easily affected by the fault condition and environmental noise, a new line selection method for small current grounding fault based on Gramian angular field (GAF) and improved residual network (PrResNet) is proposed. Firstly, a simulation model is built to collect current data, and the one- dimensional zero-sequence current timing information is converted into two-dimensional image information by GAF transform. Then, residual network (ResNet) is improved and PrResNet is proposed to extract image features. Finally, line selection results are output through Softmax. By testing the effect of GAF-PrResNet on line selection under different operating conditions and fault conditions of distribution network, it is shown that the fault line selection model based on GAF-PrResNet can realize high-precision line selection under different fault locations and transition resistance conditions under strong noise interference scenes, and has strong robustness.

Key words： small current grounding distribution network grounding fault line selection Gramian angular field (GAF) transform improved residual network (PrResNet)

收稿日期: 2023-02-27

基金资助:海南省科技专项（ZDYF2022SHFZ047）

作者简介: 刘伟（1971—）,男,黑龙江省宾县人,教授,博士生导师,研究方向为智能监测与诊断系统、电力系统与智能控制。

引用本文:

刘伟, 杨东风, 王洪志, 闫文迪. 基于格拉姆角场-改进残差网络的小电流接地系统故障选线[J]. 电气技术, 2023, 24(12): 14-19. LIU Wei, YANG Dongfeng, WANG Hongzhi, YAN Wendi. Fault line selection method for small current grounding system based on Gramian angular field-improved residual network. Electrical Engineering, 2023, 24(12): 14-19.

链接本文:

http://dqjs.cesmedia.cn/CN/Y2023/V24/I12/14

[1] 闫严. 基于EMD与FIR滤波的小电流接地故障选线方法[J]. 电气技术, 2017, 18(4): 56-60.
[2] 杨帆, 牟龙华, 张鑫, 等. 基于序分量差异性的小电流接地系统单相接地故障选线法[J]. 电气技术, 2019, 20(3): 13-17, 23.
[3] 邓丰, 梅龙军, 唐欣, 等. 基于时频域行波全景波形的配电网故障选线方法[J]. 电工技术学报, 2021, 36(13): 2861-2870.
[4] 张淑清, 马跃, 李盼, 等. 基于改进的广义谐波小波包分解和混沌振子的小电流接地系统故障选线[J]. 电工技术学报, 2015, 30(3): 13-20, 43.
[5] WANG Yuanyuan, ZENG Xiangjun, DONG Zhaoyang, et al.Stator single-phase-to-ground fault protection for bus-connected powerformers based on hierarchical clustering algorithm[J]. IEEE Transactions on Energy Conversion, 2013, 28(4): 991-998.
[6] 方毅, 薛永端, 宋华茂, 等. 谐振接地系统高阻接地故障暂态能量分析与选线[J]. 中国电机工程学报, 2018, 38(19): 5636-5645.
[7] 张祚淏, 王利恒. 基于小波分析和GA-SVM的小电流接地故障选线方法[J]. 自动化与仪表, 2021, 36(8): 18-23.
[8] 韦莉珊, 贾文超, 焦彦军. 基于5次谐波与导纳不对称度的配电网单相接地选线方法[J]. 电力系统保护与控制, 2020, 48(15): 77-83.
[9] 薛太林, 侯隽朗, 张建新. 基于GA优化BP神经网络小电流接地系统故障选线方法[J]. 电气自动化, 2018, 40(2): 66-69.
[10] 陈奎, 陈博博. 基于改进暂态相关分析和支持向量机的电弧故障选线研究[J]. 电力系统保护与控制, 2016, 44(24): 66-73.
[11] 高金峰, 秦瑜瑞, 殷红德. 基于小波包变换和支持向量机的故障选线方法[J]. 郑州大学学报(工学版), 2020, 41(1): 63-69.
[12] 孟安波, 葛佳菲, 李德强, 等. 基于纵横交叉算法的神经网络配电网故障选线研究[J]. 电力系统保护与控制, 2016, 44(21): 90-95.
[13] GUO Moufa, ZENG Xiaodan, CHEN Duanyu, et al.Deep-learning-based earth fault detection using conti- nuous wavelet transform and convolutional neural network in resonant grounding distribution systems[J]. IEEE Sensors Journal, 2017, 18(3): 1291-1300.
[14] JI Tao, PANG Qingle, LIU Xinyun.Study on fault line detection based on genetic artificial neural network in compensated distribution system[C]//2006 IEEE Inter- national Conference on Information Acquisition, Weihai, China, 2007.
[15] 郝帅, 马瑞泽, 赵新生, 等. 基于卷积块注意模型的YOLOv3输电线路故障检测方法[J]. 电网技术, 2021, 45(8): 2979-2987.
[16] 郝帅, 张旭, 马瑞泽, 等. 基于改进GoogLeNet的小电流接地系统故障选线方法[J]. 电网技术, 2022, 46(1):361-368.
[17] HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 2016.