摘要
针对电力系统出现的小电流接地故障,提出一种利用BP神经网络对暂态主频分析的选线方法.该方法提取各出线初始零模电流暂态主频的实部、虚部,利用BP神经网络实现故障线路的选取.使用建模软件ATP-EMTP建立小电流接地故障模型,对各条线路分别设置不相同故障状态(故障距离、过渡电阻和故障初相角)进行仿真,得到BP神经网络所需训练样本和测试数据,利用MATLAB进行编程实现神经网络的学习与训练,实现非线性映射.再利用训练好的神经网络模型对小电流故障接地故障进行选线.仿真结果表明,采用本方法可以很好地完成小电流故障接地系统的选线.
For the non-effectively earthed system, this paper propose a method of using neural network to solve the problem of fault line selection. This method extract real and imaginary parts of transient frequency to use the BP neural network to find the fault line. First, a simulation model of non-effectively earthed system was put up by ATP-EMTP software, to simulate different fault distance, transition resistance and fault initial angle in the model to obtain training samples and testing samples. Second, MATLAB was used to train the BP neural network model. At last, the trained neural network was used to find the fault line. The simulation proved that the method can solve the problem of line selection of the non-effectively earthed system.
作者
姜晓东
任力
刘铭
李彦
陈平
JIANG Xiao-dong REN Li LIU Ming LI Yan CHEN Ping(School of Electrical and Electronic Engineering, Shandong University of Technology, Zibo 255049, China Rizhao Power Supply Company State Grid Shandong, Rizhao 276826, China)
出处
《山东理工大学学报（自然科学版）》
CAS
2017年第1期67-70,共4页
Journal of Shandong University of Technology:Natural Science Edition
关键词
小电流接地系统
暂态
BP神经网络
故障选线
non- effectively earthed system
transient state
BP neural network
fault line selection
