摘要
为降低配电网单相接地故障智能定位误差,提出基于小波包的配电网单相接地故障智能定位方法。该方法通过对小波包原理的研究,建立了配电网中用小波包进行单相接地故障智能测距的数学模型,该模型中的隐层激励函数以小波函数为主,而且利用小波能量与暂态零序电流有功无功特征相结合的方法,实现了电力系统中的信息融合,完成了对系统故障的定位,而且在训练过程中加入了动量项,采用了一种新的权重初始化方式,提高了网络的学习效率,进而降低了故障定位误差。经MATLAB进行了仿真试验,结果表明所提出的方法是有效的,可实现较为精准的故障定位。
In order to reduce the intelligent positioning error of single-phase grounding faults in distribution networks,a wavelet packet based intelligent positioning method for single-phase grounding faults in distribution networks is proposed.This method establishes a mathematical model for intelligent location of single-phase grounding faults in distribution networks using wavelet packets through the study of the principle of wavelet packets.The hidden layer excitation function in this model is mainly wavelet function,and the method of combining wavelet energy with the active and reactive characteristics of transient zero sequence current realizes information fusion in the power system,completes the positioning of system faults,and adds momentum term during the training process,A new weight initialization method has been adopted to improve the learning efficiency of the network and thereby reduce fault location errors.Simulation experiments were conducted using MATLAB,and the results showed that the proposed method is effective and can achieve more accurate fault localization.
作者
史文正
Shi Wenzheng(Jiangsu Sudian Industry Management Co.,Ltd.,Nanjing 210024,China)
出处
《绿色建造与智能建筑》
2024年第9期171-175,共5页
Green Construction and Intelligent Building
关键词
故障智能定位
小波包
零序电流
intelligent fault localization
wavelet packet
zero sequence current