摘要
针对小电流接地系统单相接地故障定位难的问题,提出了一种利用改进的基于梯度的神经-模糊学习算法故障定位新方法。构造了一个由输入层、语言项层、规则层、输出层组成的模糊神经网络,通过该算法大大减小了过渡电阻对故障定位的影响。仿真结果表明,所提出的基于梯度的神经-模糊理论、模型及算法具有较好的故障定位性能。
For the the difficulty of location about single-phase to ground fault occurred in small current neutral grounding system,we put forward a new way to use improved neural-fuzzy learning algorithm,which is based on gradient for fault location.The article constructs a fuzzy neural network,which includes the input layer,language layer,rule layer and output layer.With the algorithm,it greatly decreases the influence of transition resistance on fault location.The simulation results show that the presented neural-fuzzy theory,which is based on the gradient,the model and the algorithm have a better performance of fault location.
出处
《仪表技术与传感器》
CSCD
北大核心
2014年第3期65-67,共3页
Instrument Technique and Sensor
基金
贵州省科学技术基金(黔科合J字[2011]2109号)
关键词
小电流接地系统
单相接地
故障定位
模糊神经网络
过渡电阻
故障点
low current earthing system
single-phase-to-earth faults
fault location
fuzzy neural work
transition resistance
fault point
