摘要
配电线路是现代电力系统的组成部分,直接影响着供电的安全和稳定。配电网故障定位分为对故障点的精准定位与区段定位两种。针对配电网结构的复杂性,提出基于GAF-ResNet50的配电网故障区段定位方法。该方法通过格拉姆角场算法,将一维时间序列转换成二维(Gramian angular field,GAF)图像,并利用残差神经网络的框架,从GAF图像中提取信号更深层次的故障特征,精确地辨识故障区域。为验证该方法的有效性,在MATLAB平台上搭建IEEE 13节点的配电网模型,对其产生故障数据进行故障区段的定位仿真。研究结果表明:该方法能够快速、准确地进行故障区段定位,其精度在98%以上,且该方法对噪声具有良好的鲁棒性。
Distribution lines are an integral part of modern power system,which directly influence the safety and stability of power supply.Distribution network fault location can be classified into precise fault location and fault segment location.Considering the complexity of distribution network structure,this paper proposes a fault segment location method based on Gramian angular field(GAF)-ResNet50.The one-dimensional time series is converted into a two-dimensional GAF image by the GAF algorithm,and the deeper-level fault features of the signal are extracted from the GAF image by using the framework of residual neural network so that the fault areas can be identified more accurately.To verify the effectiveness of the proposed method,the study builds an IEEE 13-node distribution network model on the MATLAB platform to generate fault data and conduct the simulation of fault segment location.The simulation results show that the proposed method can quickly and accurately locate fault segments,with a positioning accuracy of more than 98%,and has good robustness to noise.
作者
石昱烜
席燕辉
张伟杰
SHI Yuxuan;XI Yanhui;ZHANG Weijie(School of Electrical&Information Engineering,Changsha University of Science&Technology,Changsha 410014,China;School of Electrical Engineering,Shaoyang University,Shaoyang 422099,China)
出处
《电力科学与技术学报》
北大核心
2025年第2期122-130,149,共10页
Journal of Electric Power Science And Technology
基金
国家自然科学基金(52277078)
湖南省自然科学基金(2022JJ30609)。
关键词
残差神经网络
配电网
格拉姆角场
域变换
故障定位
residual neural network
distribution network
Gramian angular field
domain transformation
fault location
