摘要
对径向基函数(RBF)神经网络进行了理论分析,采用基于熵的模糊聚类(EFC)算法确定径向基神经网络的隐含层中心矢量,形成了EFC-RBF神经网络算法,并尝试将该算法用于变压器故障模式识别.仿真实验表明:EFC-RBF神经网络算法的训练过程比反向传播(BP)神经网络表现更优,实际故障数据验证其能够进行变压器故障模式识别,具有故障诊断的有效性.
Based on analyzing radial basis function(RBF)neural network,using the entropy of fuzzy clustering(EFC)algorithm,we determined its hidden centric vector.Then the EFC-RBF neural network algorithm was formed and used to transformer fault pattern recognition.Simulation results showed that EFC-RBF neural network algorithm has better performance than the back-propagation neural network in training,and the actual fault data can prove that this algorithm is applicable to fault pattern recognition.
出处
《云南大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第S2期182-186,共5页
Journal of Yunnan University(Natural Sciences Edition)
关键词
径向基函数神经网络
变压器
故障模式识别
熵
模糊聚类
radial basis function neural network
power transformer
fault pattern recognition
entropy
fuzzy clustering
