摘要
输电线路故障受到许多不确定性因素影响,传统测距算法不能很好地解决输电线路测距问题。提出了基于模糊神经网络(FNN)的故障测距方法,构造了一个由输入层、模糊化层、推理层、去模糊化层、输出层组成的模糊神经网络,采用变步长方法改进BP算法,加快了收敛速度。仿真结果显示,设计的网络具有良好的适应性能,其测距精度较高,且测距精度不受系统运行方式、过渡电阻、两端系统相角差等的影响。
As affected by many uncertain factors of transmission lines ,conventional fault locating methods can＇t locate the faults of transmission lines very well. A fault locating method based fuzzy neural network is proposed,which consists of layers of input,fuzzy,reasoning,de-fuzzy and output layers. It adopts variable steps to improve BP algorithm and quicken convergence speed. Simulation results show that it has good adaptab operation modes, transition resistances ility and high precision,and is not affected by different system and phase angle difference of two terminals.
出处
《电力自动化设备》
EI
CSCD
北大核心
2006年第5期32-34,74,共4页
Electric Power Automation Equipment
关键词
输电线路
故障测距
模糊神经网络
分层分布
transmission line
fault locating
fuzzy neural network
layered distribution
